From 2fcef136eb2fbbed1a5b99cf838afd21b00a9d6c Mon Sep 17 00:00:00 2001
From: lars20070 <l.nilse@dunelm.org.uk>
Date: Fri, 31 Jul 2026 08:27:55 +0200
Subject: [PATCH 01/15] Add Cursor config

---
 .cursor/mcp.json      | 17 +++++++++++++++++
 .cursor/settings.json | 11 +++++++++++
 2 files changed, 28 insertions(+)
 create mode 100644 .cursor/mcp.json
 create mode 100644 .cursor/settings.json

diff --git a/.cursor/mcp.json b/.cursor/mcp.json
new file mode 100644
index 0000000..142745e
--- /dev/null
+++ b/.cursor/mcp.json
@@ -0,0 +1,17 @@
+{
+  "mcpServers": {
+    "context7": {
+      "command": "npx",
+      "args": [
+        "-y",
+        "@upstash/context7-mcp@latest"
+      ]
+    },
+    "github": {
+      "url": "https://api.githubcopilot.com/mcp/",
+      "headers": {
+        "Authorization": "Bearer ${env:GITHUB_TOKEN}"
+      }
+    }
+  }
+}
\ No newline at end of file
diff --git a/.cursor/settings.json b/.cursor/settings.json
new file mode 100644
index 0000000..0fc8894
--- /dev/null
+++ b/.cursor/settings.json
@@ -0,0 +1,11 @@
+{
+  "mcpServers": {
+    "context7": {
+      "disabled": false
+    },
+    "github": {
+      "disabled": false
+    }
+  },
+  "remote.containers.reopenFolderInContainer": false
+}
\ No newline at end of file

From fb680e6dddfac6736955845b3996ff94189b88a9 Mon Sep 17 00:00:00 2001
From: lars20070 <l.nilse@dunelm.org.uk>
Date: Fri, 31 Jul 2026 10:31:39 +0200
Subject: [PATCH 02/15] Plan for scraping Google style guide

---
 .../plans/google_style_guide_scraper_plan.md  | 313 ++++++++++++++++++
 1 file changed, 313 insertions(+)
 create mode 100644 .cursor/plans/google_style_guide_scraper_plan.md

diff --git a/.cursor/plans/google_style_guide_scraper_plan.md b/.cursor/plans/google_style_guide_scraper_plan.md
new file mode 100644
index 0000000..67e7214
--- /dev/null
+++ b/.cursor/plans/google_style_guide_scraper_plan.md
@@ -0,0 +1,313 @@
+# Coding Agent Specification: Google Style Guide Single-File Markdown Generator
+
+## 1. Problem Statement
+
+### 1.1 Overview
+The **Google Developer Documentation Style Guide** (hosted at `https://developers.google.com/style`) is an essential reference for technical writers, software engineers, and AI prompt engineers. However, Google publishes this guide across **hundreds of discrete web pages** without providing an official single-file download (such as a consolidated PDF or Markdown document).
+
+### 1.2 Target Audience & Use Case
+A single-file Markdown (`.md`) document is required to:
+1. Provide **offline accessibility** and fast local text searching.
+2. Serve as a **clean context document** for Large Language Models (LLMs) and RAG (Retrieval-Augmented Generation) pipelines.
+3. Enable easy version-controlled local archival and custom printing.
+
+### 1.3 Technical Challenges
+- **Navigation Hierarchy:** Page order matters. The scraper must parse the left-hand navigation sidebar to retain the intended reading sequence.
+- **Boilerplate & Noise Removal:** Google DevSite pages contain extensive non-content UI (navigation bars, search boxes, header/footer, rating widgets, sidebars, cookie banners).
+- **Custom DOM Elements:** DevSite uses custom tags and CSS classes for callouts (`aside class="note"`, `aside class="caution"`), custom code tabs, and structured tables.
+- **Internal Link Disruption:** Relative links pointing to `https://developers.google.com/style/...` will break unless converted into local section anchors (`#...`).
+- **Rate Limiting & Politeness:** Fetching hundreds of pages sequentially requires request throttling, proper User-Agent headers, and retry logic.
+
+---
+
+## 2. Solutions Overview
+
+### Option Comparison
+
+| Approach | Pros | Cons | Decision |
+|---|---|---|---|
+| **A. Static HTML Scraper (`httpx` + `BeautifulSoup` + `markdownify`)** | Fast, lightweight, pure Python, highly customizable AST conversion. | Requires manually mapping custom DOM elements. | **RECOMMENDED** |
+| **B. Headless Browser (`Playwright` + Node/Python + Turndown)** | Renders JavaScript-heavy elements automatically. | Slower, heavier resource usage, unnecessary overhead for static text pages. | Secondary fallback |
+| **C. Shell Pipeline (`wget` + `pandoc`)** | Simple one-liner execution. | Messy output, preserves UI clutter, mangles navigation order, poor callout handling. | Rejected |
+
+### Recommended Solution Architecture
+A modular Python script operating in 6 sequential stages:
+1. **Discover:** Extract page hierarchy from sidebar navigation.
+2. **Fetch:** Retriable HTTP fetches with concurrent batching and caching.
+3. **Clean:** Target main content containers and strip site chrome/UI.
+4. **Transform:** Convert HTML AST to clean Markdown with custom rules for code and callouts.
+5. **Normalize:** Update internal cross-references to point to document anchors.
+6. **Compile:** Assemble master TOC, frontmatter, and output single `.md` file.
+
+---
+
+## 3. Detailed Staged Implementation Plan
+
+```
+[ Stage 1: Setup ] ──> [ Stage 2: Crawl Nav Tree ] ──> [ Stage 3: Fetch & Clean HTML ]
+                                                                     │
+[ Stage 6: Validation ] <── [ Stage 5: Compile & Normalize ] <── [ Stage 4: HTML -> MD ]
+```
+
+### Stage 1: Environment & Dependency Setup
+Set up a python environment with required libraries:
+- `httpx` (Async HTTP client with HTTP/2 and retry capabilities)
+- `beautifulsoup4` + `lxml` (Fast HTML parsing and tree navigation)
+- `markdownify` (HTML to Markdown converter with custom subclassing)
+- `pyyaml` (Frontmatter generation)
+
+#### Required Directory Layout
+```text
+scraper/
+├── cache/                  # Downloaded raw HTML files (prevents refetching)
+├── output/                 # Output directory for final MD file
+├── config.py               # CSS selectors, user agents, rate limits
+├── parser.py               # Custom HTML-to-MD rules
+└── main.py                 # CLI entrypoint and orchestrator
+```
+
+---
+
+### Stage 2: Sitemap & Hierarchy Discovery
+**Objective:** Parse the left navigation menu on `https://developers.google.com/style` to build an ordered list of URLs with section levels.
+
+#### Step-by-Step Instructions:
+1. Fetch the main index page `https://developers.google.com/style`.
+2. Locate the navigation tree element:
+   - Primary selector: `ul.devsite-nav-section-list` or `nav.devsite-section-nav`.
+3. Extract all `<a>` tags with `href` attributes starting with `/style`.
+4. Store URLs sequentially in a list of structured dicts:
+   ```python
+   page_item = {
+       "title": "Voice and tone",
+       "url": "https://developers.google.com/style/voice",
+       "level": 2,  # Depth in nav hierarchy
+       "slug": "voice"
+   }
+   ```
+5. Deduplicate links while strictly maintaining first-seen order.
+
+---
+
+### Stage 3: Fetching & DOM Cleaning
+**Objective:** Fetch raw HTML for each page, store in cache, and strip non-article elements.
+
+#### Step-by-Step Instructions:
+1. **Caching:** Check if `cache/{slug}.html` exists. If not, fetch via HTTP request with `User-Agent: Mozilla/5.0 ...`.
+2. **Rate Limiting:** Implement a 200–500ms delay between requests to avoid 429 throttling.
+3. **Locate Core Content:**
+   - Target container: `article.devsite-article` or `div.devsite-article-body`.
+4. **De-cluttering (Removal List):**
+   Remove the following selectors before processing:
+   - `nav`, `header`, `footer`
+   - `.devsite-rating-container` (Feedback widgets)
+   - `.devsite-toc` (In-page right-hand TOC)
+   - `.devsite-content-footer`
+   - `script`, `style`, `noscript`
+   - Buttons like "Copy code" or print triggers.
+
+---
+
+### Stage 4: Custom HTML-to-Markdown Conversion
+**Objective:** Transform cleaned HTML DOM into clean Markdown while handling Google-specific elements.
+
+#### Custom Tag Handling Rules:
+
+1. **Headings (`<h1>` to `<h6>`):**
+   - Shift header ranks down by 1 level if using `# Title` for the main page document.
+   - Inject HTML anchor IDs or explicit slug targets: `<a id="slug-heading"></a>`.
+
+2. **Callout Boxes / Notices (`<aside>`):**
+   Google uses `<aside class="note">`, `<aside class="caution">`, `<aside class="key-point">`.
+   - Convert to GitHub-Flavored Markdown (GFM) admonition syntax:
+     ```markdown
+     > [!NOTE]
+     > Note text content here...
+     ```
+
+3. **Code Blocks (`<pre>`, `<code>`):**
+   - Extract language class if present (e.g. `class="lang-py"` -> `py`).
+   - Wrap in triple backticks ` ```python ... ``` `.
+
+4. **Tables (`<table>`):**
+   - Convert standard `<table>` / `<tr>` / `<th>` / `<td>` to Markdown pipe tables.
+   - Clean up interior multiline HTML inside table cells to single-line or `<br>`.
+
+---
+
+### Stage 5: Link Normalization & Single-File Compilation
+**Objective:** Assemble all parsed sections into a single monolithic `.md` file and fix internal hyperlinks.
+
+#### Step-by-Step Instructions:
+1. **Frontmatter:** Inject YAML frontmatter at top of document:
+   ```yaml
+   ---
+   title: Google Developer Documentation Style Guide
+   source: https://developers.google.com/style
+   generated_at: YYYY-MM-DD
+   format: Single-file Markdown
+   ---
+   ```
+2. **Master Table of Contents:**
+   Generate a master TOC linking to every page heading using synthesized anchors (e.g. `#voice-and-tone`).
+3. **Link Rewriting Rule:**
+   - External links (`https://...` outside `/style`) -> Keep as-is.
+   - Internal links (`/style/voice` or `https://developers.google.com/style/voice`) -> Convert to local Markdown anchor links `[Voice and tone](#voice)`.
+4. **Page Dividers:**
+   Insert page section headers and horizontal rules between scraped pages:
+   ```markdown
+   ---
+   # Page Title
+   *Source: https://developers.google.com/style/page*
+   
+   [Page Content]
+   ```
+
+---
+
+### Stage 6: Validation & Verification
+**Objective:** Guarantee the quality and completeness of the generated file.
+
+#### Automated Checks to Implement in Script:
+1. **Missing Content Check:** Ensure page count in output matches total discovered URLs.
+2. **Unbroken Links Check:** Verify all internal `#anchor` links point to valid targets within the generated document.
+3. **File Size Check:** Expected size is typically between 1.5 MB and 4.0 MB.
+4. **Syntax Check:** Ensure code blocks and tables do not break Markdown rendering syntax.
+
+---
+
+## 4. Ready-to-Run Reference Implementation Script
+
+Below is a complete, runnable Python script that implements the above specification. A coding agent can execute or extend this directly:
+
+```python
+import os
+import re
+import time
+from urllib.parse import urljoin, urlparse
+import httpx
+from bs4 import BeautifulSoup
+from markdownify import MarkdownConverter
+
+BASE_URL = "https://developers.google.com/style"
+OUTPUT_FILE = "google_developer_style_guide.md"
+HEADERS = {"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64)"}
+
+class GoogleStyleGuideConverter(MarkdownConverter):
+    def convert_aside(self, el, text, convert_as_inline):
+        classes = el.get("class", [])
+        kind = "NOTE"
+        if "caution" in classes or "warning" in classes:
+            kind = "WARNING"
+        elif "key-point" in classes:
+            kind = "IMPORTANT"
+        
+        lines = text.strip().split("
+")
+        quoted = "
+> ".join(lines)
+        return f"
+
+> [!{kind}]
+> {quoted}
+
+"
+
+def get_navigation_urls():
+    print("Fetching navigation index...")
+    res = httpx.get(BASE_URL, headers=HEADERS, follow_redirects=True)
+    soup = BeautifulSoup(res.text, "lxml")
+    
+    urls = []
+    # Find navigation items
+    nav = soup.find("ul", class_="devsite-nav-section-list") or soup
+    for a in nav.find_all("a", href=True):
+        href = a["href"]
+        if href.startswith("/style") or href.startswith(BASE_URL):
+            full_url = urljoin(BASE_URL, href).split("#")[0]
+            if full_url not in urls:
+                urls.append(full_url)
+    print(f"Discovered {len(urls)} pages in navigation.")
+    return urls
+
+def clean_and_convert_html(html_content, page_url):
+    soup = BeautifulSoup(html_content, "lxml")
+    
+    # Extract main article
+    article = soup.find("article", class_="devsite-article") or soup.find("div", class_="devsite-article-body")
+    if not article:
+        article = soup.find("main") or soup.body
+
+    # Strip junk elements
+    for selector in [
+        "nav", ".devsite-rating-container", ".devsite-toc",
+        ".devsite-content-footer", "script", "style", "iframe"
+    ]:
+        for el in article.select(selector):
+            el.decompose()
+
+    # Rewrite internal links to local anchors
+    for a in article.find_all("a", href=True):
+        href = a["href"]
+        if href.startswith("/style") or href.startswith(BASE_URL):
+            parsed = urlparse(href)
+            slug = parsed.path.strip("/").replace("/", "-")
+            a["href"] = f"#{slug}" if slug else "#introduction"
+
+    converter = GoogleStyleGuideConverter(heading_style="ATX")
+    return converter.convert(str(article))
+
+def build_single_file_md():
+    urls = get_navigation_urls()
+    if not urls:
+        urls = [BASE_URL] # Fallback
+        
+    md_sections = []
+    md_sections.append("---
+title: Google Developer Documentation Style Guide
+source: https://developers.google.com/style
+---
+
+")
+    md_sections.append("# Google Developer Documentation Style Guide
+
+")
+
+    for idx, url in enumerate(urls, start=1):
+        parsed = urlparse(url)
+        slug = parsed.path.strip("/").replace("/", "-") or "introduction"
+        print(f"[{idx}/{len(urls)}] Processing: {url}")
+        
+        try:
+            res = httpx.get(url, headers=HEADERS, follow_redirects=True)
+            if res.status_code == 200:
+                md_content = clean_and_convert_html(res.text, url)
+                md_sections.append(f"
+
+<a id="{slug}"></a>
+
+---
+
+")
+                md_sections.append(md_content)
+            time.sleep(0.2)
+        except Exception as e:
+            print(f"Error fetching {url}: {e}")
+
+    with open(OUTPUT_FILE, "w", encoding="utf-8") as f:
+        f.writelines(md_sections)
+    print(f"Successfully generated single-file style guide: {OUTPUT_FILE}")
+
+if __name__ == "__main__":
+    build_single_file_md()
+```
+
+---
+
+## 5. Summary Checklist for Agent Execution
+
+- [ ] Execute script to discover navigation links.
+- [ ] Verify HTML sanitization rules strip header, footer, and navigation.
+- [ ] Confirm custom `<aside>` conversion produces GFM callout blockquotes.
+- [ ] Ensure internal links are mapped to `#slug` local anchors.
+- [ ] Validate final single `.md` file size and content integrity.

From 5a64678a6765a5a49cba45e1812077a2a42bf4a7 Mon Sep 17 00:00:00 2001
From: lars20070 <l.nilse@dunelm.org.uk>
Date: Fri, 31 Jul 2026 10:34:17 +0200
Subject: [PATCH 03/15] New plan

---
 .../plans/google-style-guide-to-markdown.md   | 301 ++++++++++++++++++
 1 file changed, 301 insertions(+)
 create mode 100644 .cursor/plans/google-style-guide-to-markdown.md

diff --git a/.cursor/plans/google-style-guide-to-markdown.md b/.cursor/plans/google-style-guide-to-markdown.md
new file mode 100644
index 0000000..ad84f37
--- /dev/null
+++ b/.cursor/plans/google-style-guide-to-markdown.md
@@ -0,0 +1,301 @@
+# Google developer documentation style guide as one Markdown file
+
+## Context
+
+`md/` wants clean, structured Markdown. The Google developer documentation
+style guide (<https://developers.google.com/style>) is published as a DevSite
+book with no single-file download, so it needs a fetch-and-convert step of its
+own — a sibling to `pdf2md/`, which does the same job for PDFs. The output is
+one file in `md/`, and from there the existing pipeline takes over: `make
+wiki-sandbox` runs Pi once per `md/*.md` file and folds it into `okf/`.
+
+This supersedes
+[.cursor/plans/google_style_guide_scraper_plan.md](.cursor/plans/google_style_guide_scraper_plan.md).
+Its overall shape — discover, fetch, clean, convert, normalize links, assemble,
+validate — is right and is kept. Its facts about the site are not, and its
+reference script would not run. Both are corrected below against measurements
+taken from the live site on 2026-07-31.
+
+## What the site actually is (measured)
+
+Ten pages were fetched and analysed: the landing page, `word-list`, `tone`,
+`notices`, `lists`, `tables`, `code-samples`, `code-syntax`, `highlights`.
+
+- **70 pages, not "hundreds."** The book nav is server-rendered in the HTML of
+  every page: 71 `/style` anchors (one is the "Guides" tab, pointing at `/style`
+  a second time) inside `.devsite-book-nav-wrapper`, grouped by 10
+  `li.devsite-nav-heading` labels — Introduction, Key resources, General
+  principles, Language and grammar, Punctuation, Formatting and organization,
+  Linking, Computer interfaces, HTML and CSS, Names and naming. No JavaScript,
+  no lazy-loaded subtree: one fetch yields the whole ordered outline.
+- **`robots.txt` allows it.** `User-agent: *` with a single `Disallow:
+  /youtube/partner/`. A truthful, non-browser User-Agent is served normally
+  (checked: `200`), so there is no reason to spoof Mozilla as the old plan did.
+- **Roughly 1.2 MB of article HTML.** Article bodies measured 5–33 KB each,
+  except `word-list` at 251 KB. Extrapolated: ~1.2 MB of body HTML, which
+  converts to about **0.5–0.9 MB of Markdown** — not the 1.5–4 MB the old plan
+  asserts as its size check.
+- **No shortcut exists.** `?print=1` returns the same single page (byte-for-byte
+  the same size), `/style/_book.yaml`, `/style/_toc.yaml`, `/llms.txt` and
+  `.txt` suffixes all 404, and `.md` / `?format=markdown` return the HTML page
+  with `Content-Type: text/html`. The `<devsite-llm-tools>` element on each page
+  is an empty JS-driven widget with no server endpoint behind it. Scraping HTML
+  is the only route.
+- **Every page carries an AI-generated summary inside the article body.**
+  `<devsite-key-takeaways-panel>` ("Page Summary", tooltip "Generated with AI")
+  sits inside `div.devsite-article-body` on all 10 sampled pages. It is not
+  source content and must be stripped — the old plan's removal list misses it,
+  along with `<devsite-recommendations>`, `<devsite-feedback>` and
+  `<devsite-thumb-rating>`, which are also inside or adjacent to the body.
+- **90% of internal links carry a fragment.** 838 of 936 `/style` links in the
+  sampled bodies point at `#anchor` targets. The old plan strips fragments and
+  maps every link to a bare page slug, which throws away almost all of the
+  cross-referencing the guide is built on.
+- **Anchor ids collide across pages.** `examples` occurs on 3 sampled pages;
+  `politeness-and-use-of-please`, `list-or-table`,
+  `some-things-to-avoid-where-possible` and
+  `some-techniques-and-approaches-to-consider` each on 2. Merging 70 pages into
+  one file therefore needs namespaced anchors, not the raw ids.
+- **The word list is a definition list, and it is 40% of the guide.** 26 `<dl>`
+  blocks, 598 `<dt>` terms, every one with an `id` (only 2 are not slug-safe:
+  `+` and `google-play-services-SDK`). markdownify 1.2.3 does convert `dt`/`dd`,
+  but into Pandoc/PHP-Markdown-Extra syntax (`term` then `:   definition`,
+  source lines 521–556) that GitHub and most renderers do not support.
+- **Some meaning exists only in CSS.** Counted in the sample:
+  `<span class="icon-dontuse">` 148 times, `icon-avoid` 93, plus `icon-android`
+  27, `icon-cloud` 9, `icon-workspace` 3. These spans are **empty** — a naive
+  conversion silently drops "don't use this term" from hundreds of word-list
+  entries. By contrast `compare-better` (158) and `compare-worse` (91) wrap the
+  words "Recommended" / "Not recommended", so those survive on their own.
+- **`material-icons` leaks the word "link" 600+ times.** Each `<dt>` ends with
+  `<a href="#id"><span class="material-icons" aria-hidden="true">link</span></a>`;
+  `word-list` alone has 603 occurrences.
+- **Code blocks have no language.** Every one is
+  `<devsite-code><pre class="devsite-click-to-copy">…</pre></devsite-code>`, and
+  no `lang-*` or `language-*` class appears anywhere in the sample. The old
+  plan's language extraction has nothing to extract; plain fences are correct.
+- **Body headings run h2 to h4 only** (60 h2, 37 h3, 5 h4 in the sample), and
+  there are 10 images, all with site-relative `src`.
+
+## Where the old reference script breaks
+
+- `convert_aside(self, el, text, convert_as_inline)` is the pre-1.0 markdownify
+  signature. Since 1.0 the hook is `convert_<tag>(self, el, text, parent_tags)`
+  (markdownify is at 1.2.3), so the method would never be called with the
+  arguments it expects.
+- Several string literals contain raw newlines inside non-triple quotes, and an
+  f-string embeds unescaped `"` — the file as written is not valid Python.
+- It declares async, concurrency, caching and retries in the prose, then fetches
+  with a bare synchronous `httpx.get` in a loop, with no cache and no retry.
+- `nav = soup.find("ul", class_="devsite-nav-section-list") or soup` — that class
+  does not exist on the page, so the fallback silently harvests every `/style`
+  link in the whole document, in DOM order rather than nav order, and loses the
+  10 section groups entirely.
+- Nothing removes the AI summary panel, and nothing handles `dl`/`dt`/`dd`, the
+  empty semantic spans, or the icon text.
+
+## `marker` and LLMs: no for converting, yes for checking
+
+`marker` is the wrong tool here, and the reason is structural, not a matter of
+tuning. Its job is to **recover** structure a PDF has already lost — reading
+order, table geometry, headings inferred from font size — and it leans on a
+language model to guess well. This source has lost nothing: DevSite ships
+semantic HTML carrying the exact `id` attributes that 838 cross-links depend on.
+Routing HTML through a headless-browser PDF print and then `marker` would
+destroy every `href` and anchor, turn a 598-term definition list into guessed
+headings or tables, cost one LLM call per page, and produce a different file on
+every run — so `git diff` on a refresh would be unreadable. Deterministic HTML
+parsing is lossless, free and fast (~25 s for 70 pages).
+
+What `pdf2md` does contribute is its **checking** chain, which applies verbatim:
+`prettier --check`, `markdownlint-cli2` and `cspell` over the generated file, as
+documented in [pdf2md/README.md](pdf2md/README.md). Reuse it rather than
+inventing a new "syntax check."
+
+An LLM is still worth one optional, off-by-default pass: `--audit N` samples N
+pages, hands the model the cleaned HTML text beside the emitted Markdown, and
+asks what content disappeared. That is exactly the class of bug the CSS-only
+`icon-dontuse` spans represent, and it is cheap on a sample. It must stay out of
+the write path so the output stays reproducible.
+
+The genuinely useful LLM stage is the one that already exists downstream:
+`make wiki-sandbox` folds the file into OKF. Keeping the scraper deterministic
+leaves that as the only nondeterministic stage in the chain. One caveat worth
+watching: a ~0.6 MB source is roughly 150K tokens for a single Pi run — fine on
+the 1M-context catalogue models the README describes, but the first run is the
+place to confirm it.
+
+## Decisions
+
+- **One output file**, `md/GoogleDeveloperDocumentationStyleGuide.md`, matching
+  the existing `md/TheEconomistStyleGuide2023.md` naming.
+- **New `web2md/` directory**, mirroring `pdf2md/`: one Python module plus a
+  README, wired into `make lint` and a `make` target.
+- **Sequential fetch with an on-disk cache.** 70 pages at ~200 ms is ~25 s; the
+  old plan's concurrency is complexity without a payoff, and a cache makes
+  re-runs of the conversion free.
+- **Deterministic output.** Same input, same bytes, so a refresh diffs cleanly.
+- **Scraper dependencies in their own dependency group**, so CI's
+  `uv run --only-group dev ruff` stays untouched and nobody needs `marker-pdf`
+  (and torch) to scrape a website.
+
+## Changes
+
+### `web2md/google_style_guide.py` (new)
+
+One module, argparse CLI, six stages. DevSite-specific selectors live in
+constants at the top so pointing it at another DevSite book is a config change.
+
+```python
+BASE = "https://developers.google.com"
+BOOK = f"{BASE}/style"
+LOCALE = {"hl": "en"}          # pin the locale; the page honours ?hl=en
+UA = "md2okf-web2md/0.1 (+https://github.com/<owner>/md2okf)"
+NAV = ".devsite-book-nav-wrapper"
+BODY = "div.devsite-article-body"
+DROP = (
+    "devsite-key-takeaways-panel", "devsite-recommendations",
+    "devsite-feedback", "devsite-thumb-rating", "devsite-toc",
+    ".devsite-rating-container", ".devsite-content-footer",
+    ".material-icons", "[aria-hidden=true]",
+    "script", "style", "noscript", "iframe",
+)
+```
+
+1. **Discover.** Fetch `BOOK`, walk `li` elements under `NAV` in document
+   order; `li.devsite-nav-heading` sets the current section, every other
+   `li.devsite-nav-item` appends `{section, title, slug, url}`. Dedupe by URL,
+   first-seen order. Assert `60 <= len(pages) <= 200` and `len(sections) >= 5`,
+   and fail loudly otherwise — a silent 3-page run is the worst outcome.
+2. **Fetch.** One `httpx.Client(headers={"User-Agent": UA},
+   follow_redirects=True)`, `params=LOCALE`, ~200 ms between requests. Cache to
+   `web2md/cache/{slug}.html`; only `--refresh` re-fetches. Retry 429 and 5xx
+   with exponential backoff in an explicit loop — note that
+   `httpx.HTTPTransport(retries=…)` retries connection errors only, not status
+   codes.
+3. **Clean.** Take `BODY`, drop everything in `DROP`, unwrap `<devsite-code>` to
+   the `<pre>` it contains. Before dropping, replace the empty semantic spans
+   with text: `icon-dontuse` → `Don't use:`, `icon-avoid` → `Avoid:`,
+   `icon-android`/`icon-cloud`/`icon-workspace` → `(Android)`/`(Cloud)`/
+   `(Workspace)`. Take the page title from the nav, not from the `h1` — the `h1`
+   lives outside the article body.
+4. **Convert.** Subclass `MarkdownConverter` (1.2.3 signature,
+   `convert_<tag>(self, el, text, parent_tags)`; hyphens in tag names become
+   underscores in method names, so `<devsite-code>` would be
+   `convert_devsite_code` if it were not unwrapped in stage 3).
+   - `convert_aside`: `note` → `> [!NOTE]`, `caution` → `> [!CAUTION]`,
+     `warning` → `> [!WARNING]`, `success` → `> [!TIP]`, anything else →
+     `> [!NOTE]` plus a warning to stderr. All four classes are confirmed
+     present on `/style/notices`.
+   - `convert_dt` / `convert_dd`: override the Pandoc-style default. Emit
+     `<a id="{anchor}"></a>**{term}**` followed by the definition as an indented
+     paragraph, so the 598 word-list ids stay reachable and the file renders on
+     GitHub.
+   - `convert_hN`: shift by +2 and **reuse DevSite's own `id`**, never a
+     synthesized slug — the 838 fragment links target those exact strings,
+     including oddities like `formatting,-punctuation,-and-organization`.
+   - `convert_img`: absolutise `src` against `BASE`, keep the alt text.
+   - `<pre>`: plain triple-backtick fences, no language tag.
+5. **Anchors and links.** Collect every `id` in every cleaned body (759 in the
+   sample — sections, headings and `dt` terms all carry them) into one map:
+   `(slug, id) -> f"{slug}--{sanitize(id)}"`, where `sanitize` maps anything
+   outside `[A-Za-z0-9._-]` to a stable replacement (2 cases in the whole guide:
+   `+` and `google-play-services-SDK`). Emit an explicit `<a id="…"></a>` at
+   each target, then rewrite:
+   - `/style/x#y` and `https://developers.google.com/style/x#y` → `#x--y`
+   - `/style/x` → `#x`
+   - bare `#y` on page `x` → `#x--y`
+   - anything else, or a `/style` target missing from the map → absolute URL,
+     logged to stderr with a count.
+6. **Assemble.** Frontmatter written directly (five static lines: `title`,
+   `source`, `generated_at`, `pages`, `format` — no `pyyaml` dependency for
+   that), then the `h1` document title, then a TOC grouped by the 10 nav
+   sections, then per page an `h2` section heading, an `h3` page title, a
+   `*Source: <url>*` line and the
+   body. Depth budget: `#` document, `##` section, `###` page, body h2→h4,
+   h3→h5, h4→h6. That lands exactly on markdownify's h6 clamp, which is why
+   validation asserts no body heading is deeper than h4.
+
+### `web2md/README.md` (new)
+
+Same shape as `pdf2md/README.md`: what it does, the commands, and the
+`prettier` / `markdownlint-cli2` / `cspell` check chain pointed at the generated
+file. State plainly that the output is a snapshot of a living document and that
+`--refresh` is how you update it.
+
+### `pyproject.toml`
+
+```toml
+[dependency-groups]
+dev = ["ruff>=0.8"]
+web2md = [
+  "httpx>=0.28.1",
+  "beautifulsoup4>=4.15.0",
+  "lxml>=6.1.1",
+  "markdownify>=1.2.3",
+]
+```
+
+Run it as `uv run --group web2md python web2md/google_style_guide.py`. `pyyaml`
+from the old plan is not needed.
+
+### `Makefile`
+
+- Add `"web2md/README.md"` to the `markdownlint-cli2` list in `lint`.
+- Add a target:
+
+```make
+# Fetch the Google developer documentation style guide into md/ as one file.
+style-guide:
+	uv run --group web2md python web2md/google_style_guide.py
+```
+
+`ruff check .` already covers the new module.
+
+### `.gitignore`
+
+Add `web2md/cache/` — raw HTML snapshots are throwaway. The generated Markdown
+needs no rule; it lands in `md/`, which is ignored already.
+
+### `.github/workflows/ci.yml` (optional)
+
+The `paths` anchor lists no `pdf2md/**` today, so a `web2md/**` entry is a
+choice rather than a fix. Add it if you want CI to lint scraper-only changes.
+
+## Validation
+
+In-script, exiting non-zero on failure:
+
+1. Pages written equals pages discovered; list any that failed.
+2. Every `#…` link target resolves to an emitted anchor; print the offenders.
+3. No `devsite-`, no `material-icons`, and no stray Pandoc definition lines (a
+   colon followed by indent) survive in the output; no body heading deeper than
+   `h4` was seen.
+4. Size between 0.4 MB and 1.2 MB, from the ~1.2 MB of measured body HTML —
+   replacing the old 1.5–4 MB guess.
+5. Word-list term count within a tolerance of the 598 measured, since a silent
+   `dl` regression there would gut 40% of the guide.
+
+Then, by hand on the host, the `pdf2md` chain:
+
+```bash
+prettier --check md/GoogleDeveloperDocumentationStyleGuide.md
+markdownlint-cli2 md/GoogleDeveloperDocumentationStyleGuide.md
+cspell md/GoogleDeveloperDocumentationStyleGuide.md
+```
+
+Finish with `make lint`, and a `make wiki-sandbox` run to confirm the file works
+as a Pi source document. `make validate` is not needed: nothing under `pi/` or
+`scripts/` changes.
+
+## Risks
+
+- **The nav is one selector deep.** Everything hinges on
+  `.devsite-book-nav-wrapper`. Hence the count assertions in stage 1: a DevSite
+  redesign should stop the run, not quietly produce a stub.
+- **The guide changes under you.** The output is a dated snapshot. `whats-new`
+  is part of the book, so the file records its own currency.
+- **CSS-only semantics may be wider than measured.** The 10-page sample found
+  five such classes. The optional `--audit` pass exists for exactly this, and a
+  cheap guard is to log every `class` attribute value dropped during cleaning,
+  so a new one shows up as a line of output rather than as missing content.

From 10e170962d36a2ec1fa7af6dc39b2ff53eb8e346 Mon Sep 17 00:00:00 2001
From: lars20070 <l.nilse@dunelm.org.uk>
Date: Fri, 31 Jul 2026 10:42:34 +0200
Subject: [PATCH 04/15] Enhance scraping plan for Google style guide

---
 .../plans/google-style-guide-to-markdown.md   | 24 +++++++++++++++----
 1 file changed, 20 insertions(+), 4 deletions(-)

diff --git a/.cursor/plans/google-style-guide-to-markdown.md b/.cursor/plans/google-style-guide-to-markdown.md
index ad84f37..344cbaa 100644
--- a/.cursor/plans/google-style-guide-to-markdown.md
+++ b/.cursor/plans/google-style-guide-to-markdown.md
@@ -28,9 +28,14 @@ Ten pages were fetched and analysed: the landing page, `word-list`, `tone`,
   principles, Language and grammar, Punctuation, Formatting and organization,
   Linking, Computer interfaces, HTML and CSS, Names and naming. No JavaScript,
   no lazy-loaded subtree: one fetch yields the whole ordered outline.
-- **`robots.txt` allows it.** `User-agent: *` with a single `Disallow:
-  /youtube/partner/`. A truthful, non-browser User-Agent is served normally
-  (checked: `200`), so there is no reason to spoof Mozilla as the old plan did.
+- **`robots.txt` allows it, and nothing blocks a scraper.** `User-agent: *` with
+  a single `Disallow: /youtube/partner/`, plus a declared sitemap. Across ~15
+  requests with three different identities — a spoofed `Mozilla/5.0`, a truthful
+  `md2okf-web2md/0.1`, and no `User-Agent` header at all — every response was
+  `200`, including six pages back to back at 0.3 s intervals. No `429`, no
+  `Retry-After`, no challenge page. So there is no reason to spoof Mozilla as
+  the old plan did. **Untested:** a sustained run across all 70 pages, and
+  repeat runs from one IP. See Risks.
 - **Roughly 1.2 MB of article HTML.** Article bodies measured 5–33 KB each,
   except `word-list` at 251 KB. Extrapolated: ~1.2 MB of body HTML, which
   converts to about **0.5–0.9 MB of Markdown** — not the 1.5–4 MB the old plan
@@ -172,7 +177,11 @@ DROP = (
    `web2md/cache/{slug}.html`; only `--refresh` re-fetches. Retry 429 and 5xx
    with exponential backoff in an explicit loop — note that
    `httpx.HTTPTransport(retries=…)` retries connection errors only, not status
-   codes.
+   codes. Two rules make throttling impossible to mistake for success, since a
+   challenge or error page still parses as HTML and would otherwise yield a
+   hollow section: **cache only `200` responses**, so a throttled reply never
+   becomes a poisoned cache entry, and **treat a missing `BODY` as a hard
+   error** that stops the run, naming the page.
 3. **Clean.** Take `BODY`, drop everything in `DROP`, unwrap `<devsite-code>` to
    the `<pre>` it contains. Before dropping, replace the empty semantic spans
    with text: `icon-dontuse` → `Don't use:`, `icon-avoid` → `Avoid:`,
@@ -293,6 +302,13 @@ as a Pi source document. `make validate` is not needed: nothing under `pi/` or
 - **The nav is one selector deep.** Everything hinges on
   `.devsite-book-nav-wrapper`. Hence the count assertions in stage 1: a DevSite
   redesign should stop the run, not quietly produce a stub.
+- **Google could throttle a run, even though nothing suggests it will.** The
+  15-request probe above was clean at three different identities, but a full
+  70-page pass has not been measured, and neither has a second pass from the
+  same IP. The stage 2 rules are the whole defence: back off on `429`, cache
+  only `200`s, and stop on a missing article body. If a run ever does trip a
+  limit, the first move is a longer delay, not a browser or a proxy — 70 pages
+  is small enough that politeness costs nothing.
 - **The guide changes under you.** The output is a dated snapshot. `whats-new`
   is part of the book, so the file records its own currency.
 - **CSS-only semantics may be wider than measured.** The 10-page sample found

From bc3ec3cbed61f39333c82eaf98edf88ec18155b8 Mon Sep 17 00:00:00 2001
From: lars20070 <l.nilse@dunelm.org.uk>
Date: Fri, 31 Jul 2026 11:16:27 +0200
Subject: [PATCH 05/15] Code for scraping Google style guide + md output

---
 .cspell.json                                 |     5 +-
 .gitignore                                   |     6 +-
 Makefile                                     |     7 +-
 md/.gitkeep                                  |     0
 md/GoogleDeveloperDocumentationStyleGuide.md | 16508 +++++++++++++++++
 pyproject.toml                               |     6 +
 uv.lock                                      |    92 +
 web2md/README.md                             |    39 +
 web2md/google_style_guide.py                 |   668 +
 9 files changed, 17327 insertions(+), 4 deletions(-)
 create mode 100644 md/.gitkeep
 create mode 100644 md/GoogleDeveloperDocumentationStyleGuide.md
 create mode 100644 web2md/README.md
 create mode 100644 web2md/google_style_guide.py

diff --git a/.cspell.json b/.cspell.json
index 557353f..4d7c709 100644
--- a/.cspell.json
+++ b/.cspell.json
@@ -27,6 +27,7 @@
     "psutil",
     "fdfind",
     "venv",
-    "okflintrc"
+    "okflintrc",
+    "pycache"
   ]
-}
+}
\ No newline at end of file
diff --git a/.gitignore b/.gitignore
index 2813865..6b2270d 100644
--- a/.gitignore
+++ b/.gitignore
@@ -7,8 +7,12 @@
 # Python
 .venv/
 
+# web2md HTML fetch cache
+web2md/cache/
+web2md/__pycache__/
+
 # md input folder
-md/*
+# md/*
 
 # okf output folder
 okf/*
diff --git a/Makefile b/Makefile
index cdcdea6..ec70f5c 100644
--- a/Makefile
+++ b/Makefile
@@ -18,7 +18,7 @@ RUFF ?= uv run ruff
 HADOLINT ?= hadolint
 
 .DEFAULT_GOAL := lint
-.PHONY: lint lint-okf validate wiki-container wiki-sandbox
+.PHONY: lint lint-okf validate wiki-container wiki-sandbox style-guide
 
 # Lint tracked Markdown, shell, Python, and the container Dockerfile. One
 # Markdown glob per runtime so deleting a runtime is a one-line removal. The
@@ -28,6 +28,7 @@ lint:
 	$(MARKDOWNLINT) \
 		"README.md" \
 		"pdf2md/README.md" \
+		"web2md/README.md" \
 		"AGENTS.md" \
 		"pi/container/agent/**/*.md" \
 		"pi/sandbox/files/home/.pi/agent/**/*.md"
@@ -56,3 +57,7 @@ wiki-container:
 # Compile the OKF wiki with the sandboxed Pi runtime (Docker Sandbox / sbx).
 wiki-sandbox:
 	./scripts/compile-wiki-sandbox.sh
+
+# Fetch the Google developer documentation style guide into md/ as one file.
+style-guide:
+	uv run --group web2md python web2md/google_style_guide.py
diff --git a/md/.gitkeep b/md/.gitkeep
new file mode 100644
index 0000000..e69de29
diff --git a/md/GoogleDeveloperDocumentationStyleGuide.md b/md/GoogleDeveloperDocumentationStyleGuide.md
new file mode 100644
index 0000000..fba07ba
--- /dev/null
+++ b/md/GoogleDeveloperDocumentationStyleGuide.md
@@ -0,0 +1,16508 @@
+---
+title: Google Developer Documentation Style Guide
+source: https://developers.google.com/style
+generated_at: 2026-07-31
+pages: 70
+format: Single-file Markdown
+---
+
+# Google Developer Documentation Style Guide
+
+*Snapshot of [https://developers.google.com/style](https://developers.google.com/style) generated 2026-07-31.*
+
+## Table of contents
+
+### Introduction
+
+- [About this guide](#style)
+- [Highlights](#highlights)
+- [What's new](#whats-new)
+- [Philosophy of this guide](#philosophy)
+
+### Key resources
+
+- [Word list](#word-list)
+- [Product names](#product-names)
+- [Text-formatting summary](#text-formatting)
+
+### General principles
+
+- [Accessibility](#accessibility)
+- [Excessive claims](#excessive-claims)
+- [Future features](#future)
+- [Global audience](#translation)
+- [Inclusive language](#inclusive-documentation)
+- [Jargon](#jargon)
+- [Prescriptive documentation](#prescriptive-documentation)
+- [Third-party content](#other-sources)
+- [Timeless documentation](#timeless-documentation)
+- [Voice and tone](#tone)
+
+### Language and grammar
+
+- [Abbreviations](#abbreviations)
+- [Active voice](#voice)
+- [Anthropomorphism](#anthropomorphism)
+- [Articles (a, an, the)](#articles)
+- [Capitalization](#capitalization)
+- [Contractions](#contractions)
+- [Pluralization](#pluralization)
+- [Possessives](#possessives)
+- [Prepositions](#prepositions)
+- [Present tense](#tense)
+- [Pronouns](#pronouns)
+- [Second person](#person)
+- [Sentence structure](#sentence-structure)
+- [Verbs in reference documents](#reference-verbs)
+
+### Punctuation
+
+- [Colons](#colons)
+- [Commas](#commas)
+- [Dashes](#dashes)
+- [Ellipses](#ellipses)
+- [Hyphens](#hyphens)
+- [Parentheses](#parentheses)
+- [Periods and end punctuation](#periods)
+- [Quotation marks](#quotation-marks)
+- [Semicolons](#semicolons)
+- [Slashes](#slashes)
+
+### Formatting and organization
+
+- [Dates and times](#dates-times)
+- [Examples](#format-examples)
+- [Figures and other images](#images)
+- [Footnotes](#footnotes)
+- [Headings and titles](#headings)
+- [Italics with terms](#italics-terms)
+- [Lists](#lists)
+- [Mathematical notation](#mathematical-notation)
+- [Notes and other notices](#notices)
+- [Numbers](#numbers)
+- [Paragraphs](#paragraph-structure)
+- [Phone numbers](#phone-numbers)
+- [Procedures](#procedures)
+- [Tables](#tables)
+- [Units of measurement](#units-of-measure)
+
+### Linking
+
+- [Cross-references and linking](#cross-references)
+- [Headings as link targets](#headings-targets)
+
+### Computer interfaces
+
+- [API reference code comments](#api-reference-comments)
+- [Code in text](#code-in-text)
+- [Code samples](#code-samples)
+- [Command-line syntax](#code-syntax)
+- [Placeholder formatting](#placeholders)
+- [UI elements and interaction](#ui-elements)
+
+### HTML and CSS
+
+- [HTML and semantic tagging](#semantic-tagging)
+- [HTML formatting](#html-formatting)
+- [Markdown versus HTML](#markdown)
+
+### Names and naming
+
+- [Example domains and names](#examples)
+- [Filenames](#filenames)
+- [Trademarks](#trademarks)
+
+## Introduction
+
+<a id="style"></a>
+
+### About this guide
+
+*Source: <https://developers.google.com/style>*
+
+This style guide provides editorial guidelines for writing clear and consistent technical
+documentation for an audience of software developers and other technical practitioners.
+
+If you're new to the guide and looking for introductory topics about our style, then start with
+[Highlights](#highlights), [Voice and tone](#tone), and
+[Text-formatting summary](#text-formatting). Otherwise, use the guide as
+a reference document for specific questions. For example, you can look up terms in the
+[word list](#word-list).
+
+<a id="style--editorial-resources"></a>
+
+#### Editorial resources
+
+We recommend using the following editorial resources.
+
+<a id="style--reference-hierarchy"></a>
+
+##### Reference hierarchy
+
+Use the following references, including this guide, in this order:
+
+1. **Project-specific style**. Follow style guidance specific to your project or product, such
+   as necessary exceptions to this guide or terms that are relevant only to your product.
+2. **This style guide**. If project-specific style guidelines don't provide explicit
+   guidance, then follow this guide.
+3. **Third-party references**. If the preceding references don't provide explicit guidance,
+   then see these third-party references, depending on the nature of your question:
+
+   | Type of question | Third-party reference |
+   | --- | --- |
+   | Spelling | Follow [Merriam-Webster.com](https://www.merriam-webster.com/). See also [Spelling](https://developers.google.com/style/spelling). |
+   | Nontechnical style | Follow [*The Chicago Manual of Style*, 17th edition](https://www.chicagomanualofstyle.org/home.html) (subscription required). |
+   | Technical style | See the [Microsoft Writing Style Guide](https://docs.microsoft.com/style-guide/welcome/). But consider whether Microsoft's guidance applies; some of it might apply only to Microsoft products and interfaces. |
+
+At multiple stages of this hierarchy, it can be helpful to look to established usage. For
+example, search your organization's documentation, or check a broad language corpus such
+as [Google Ngram Viewer](https://books.google.com/ngrams/).
+
+<a id="style--other-editorial-resources"></a>
+
+##### Other editorial resources
+
+You can use additional resources to research and inform your thinking, but don't consider them
+part of Google developer documentation style.
+
+Here are some other style guides from the tech community:
+
+- [Apple Style Guide](https://help.apple.com/applestyleguide/)
+- [Red Hat supplementary style guide for product documentation](https://redhat-documentation.github.io/supplementary-style-guide/)
+
+<a id="style--annotations"></a>
+
+<a id="style--annotations-used-in-this-guide"></a>
+
+#### Annotations used in this guide
+
+For guidance that applies only to Android or Google Cloud documentation, look for the following
+logos:
+
+- precedes terms and guidelines specific to Android
+  documentation.
+- precedes terms and guidelines specific to Google Cloud
+  documentation.
+
+<a id="style--rules"></a>
+
+<a id="style--break-the-rules"></a>
+
+#### Break the rules
+
+> *Break any of these rules sooner than say anything outright barbarous.*
+>
+> —George Orwell,
+> "[Politics and the English Language](https://www.orwellfoundation.com/the-orwell-foundation/orwell/essays-and-other-works/politics-and-the-english-language/)"
+
+This guide contains guidelines, not rules. Depart from it when doing so improves your
+content.
+
+For example, if we recommend spelling a term as one word, and you determine that the
+hyphenated version of a term in your domain is more appropriate for your readers, then
+it's fine to use that instead. We acknowledge that sometimes there are competing forms
+of the same word in wide use, especially as new terms emerge, and you might have good
+reasons for departing from our guidance.
+
+When you depart from this guide, be consistent throughout your document.
+
+---
+
+<a id="highlights"></a>
+
+### Highlights
+
+*Source: <https://developers.google.com/style/highlights>*
+
+The style guide covers a lot of material, so the following page provides an overview of its most
+important points. For more information about topics on the page, follow the links.
+
+<a id="highlights--tone"></a>
+
+<a id="highlights--tone-and-content"></a>
+
+#### Tone and content
+
+- [Be conversational and friendly](#tone) without being
+  frivolous.
+- [Don't pre-announce anything](#future) in
+  documentation.
+- [Use descriptive link text](#cross-references--descriptive-link-text).
+- [Write accessibly](#accessibility).
+- [Write for a global audience](#translation).
+
+<a id="highlights--language"></a>
+
+<a id="highlights--language-and-grammar"></a>
+
+#### Language and grammar
+
+- [Use second person](#person): "you" rather than
+  "we."
+- [Use active voice](#voice): make clear who's performing
+  the action.
+- [Use standard American spelling](https://developers.google.com/style/spelling) and
+  punctuation.
+- [Put conditions before instructions](#sentence-structure),
+  not after.
+- [For usage and spelling of specific words, see
+  the word list](https://developers.google.com/style/wordlist).
+
+<a id="highlights--formatting"></a>
+
+<a id="highlights--formattingu002c-punctuationu002c-and-organization"></a>
+
+#### Formatting, punctuation, and organization
+
+- [Use sentence case](#capitalization) for document
+  titles and section headings.
+- [Use numbered lists](#lists--types-of-lists) for sequences.
+- [Use bulleted lists](#lists--types-of-lists) for most other lists.
+- [Use description lists](#lists--types-of-lists) for pairs of related
+  pieces of data.
+- [Use serial commas](https://developers.google.com/style/commas-serial).
+- [Put code-related text in code font](#code-in-text).
+- [Put UI elements in bold](#ui-elements).
+- [Use unambiguous date formatting](#dates-times).
+
+<a id="highlights--images"></a>
+
+#### Images
+
+- [Provide alt
+  text](#images--text-associated-with-images).
+- [Provide high-resolution or vector
+  images](#images--high-resolution-images) when practical.
+
+---
+
+<a id="whats-new"></a>
+
+### What's new
+
+*Source: <https://developers.google.com/style/whats-new>*
+
+This page provides a summary of significant changes to the style guide.
+
+<a id="whats-new--20260707"></a>
+
+<a id="whats-new--july-7u002c-2026"></a>
+
+#### July 7, 2026
+
+| New guidance or change | Page |
+| --- | --- |
+| Softened a statement regarding the effect of inconsistent terminology on translation costs. | [Write for a global audience](#translation) |
+| Added cross-references between guidance about optional procedure steps and guidance about optional headings. | [Headings and titles](#headings),  [Procedures](#procedures) |
+| Clarified that much of our guidance about writing inclusive documentation relates to the broader principle of avoiding figurative language, which can be ableist or unnecessarily graphic. Instead, we use literal, precise terms in their primary sense. | [Write inclusive documentation](#inclusive-documentation),  [Voice and tone](#tone),  [Word list](#word-list) |
+| Updated guidance about creating custom heading targets, making anchor (`<a>`) elements equally as acceptable as section (`<section>`) elements. | [Make headings into link targets](#headings-targets) |
+| Added guidance about ensuring that you contextualize UI elements when you document them outside of a numbered procedure. | [UI elements and interaction](#ui-elements) |
+| Added word list entry: *managed instance group (MIG)* | [Word list](#word-list) |
+
+<a id="whats-new--20260407"></a>
+
+<a id="whats-new--april-7u002c-2026"></a>
+
+#### April 7, 2026
+
+| New guidance or change | Page |
+| --- | --- |
+| Added guidance about avoiding inconsistent end punctuation in list items. | [Lists](#lists) |
+| Added *do the following* as a recommended phrase for introducing lists in procedures. | [Procedures](#procedures) |
+| Clarified that ordered lists are appropriate for any list where sequence is significant. Added guidance about ensuring that it's clear whether the items in an unordered list are required or optional. | [Lists](#lists) |
+| Added guidance that if you must refer to a step number, use the numeral. | [Numbers](#numbers) |
+| Added guidance to use italics sparingly, and consolidated italics guidance into a new page. | [Use italics to discuss terms](#italics-terms),  [Text-formatting summary](#text-formatting) |
+| Added guidance recommending the terms *selected* and *not selected* to refer to the state of a checkbox. | [UI elements and interaction](#ui-elements) |
+| Changed guidance to recommend using code font for IP addresses and port numbers. Added package names to the list of items to place in code font. | [Code in text](#code-in-text) |
+| Updated UI elements guidance with expanded definitions for *pane*, *panel*, and *section*. Added guidance about how to identify difficult-to-find UI elements without using directional language. | [UI elements and interaction](#ui-elements) |
+| Added `.wasm` (Wasm file) to the table of file extensions and corresponding file type names. | [Filenames and file types](#filenames) |
+| Reorganized and extended guidance about formatting abbreviation introductions. | [Abbreviations](#abbreviations),  [Text-formatting summary](#text-formatting) |
+| Consolidated pluralization guidance into a new page, including guidance about abbreviations, product names, and code elements. | [Pluralization](#pluralization) |
+| Restructured guidance about writing for a global audience to make it easier to navigate. | [Write for a global audience](#translation) |
+| Expanded guidance about exclamation marks to clarify that we avoid them except in rare cases. | [Periods and other end punctuation](#periods),  [Voice and tone](#tone) |
+| Created a page about how to format common mathematical notation. | [Mathematical notation](#mathematical-notation),  [Numbers](#numbers),  [Text-formatting summary](#text-formatting),  [Units of measurement](#units-of-measure) |
+| Changed guidance for temperatures to recommend a nonbreaking space between a numeral and the degree symbol instead of between the degree symbol and the temperature scale. | [Units of measurement](#units-of-measure) |
+| Added guidance about marking headings as optional. Restructured headings guidance to make it easier to navigate. | [Headings and titles](#headings) |
+| Added an entry for *AI*, specifying that it rarely needs to be spelled out. | [Word list](#word-list),  [Abbreviations](#abbreviations) |
+| Clarified that the word *can* can be used to convey both permission and ability. | [Word list](#word-list) |
+| Clarified that when we mark a term "Use with caution" in the word list, we recommend following our standard jargon guidance. | [Word list](#word-list) |
+| Clarified how we prefer to distinguish between a *page* (the whole web page) and a *document* (the text on a page that explains a product, feature, or service). | [Word list](#word-list) |
+| Revised the entry for *style sheet* to also allow for *stylesheet*, prioritizing consistency in a document. | [Word list](#word-list) |
+| Added explanations to several existing word list items regarding compound word usage: *clickthrough*, *hardcode*, *high availability*, *load balancing*, *plugin*, *third-party*, *time zone*, and *wake lock*. | [Word list](#word-list) |
+| Expanded the entry for *first class*, *first-class*, *first-class citizen*, providing new recommended alternatives with examples. | [Word list](#word-list),  [Write inclusive documentation](#inclusive-documentation) |
+| Expanded or added entries for *like*, *such as*, *for example*, and *for instance*. Made corresponding updates to the page about writing examples. | [Word list](#word-list),  [Format examples](#format-examples) |
+| Added guidance to the *virtual machine (VM) instance* entry regarding Compute Engine instances. | [Word list](#word-list) |
+
+<a id="whats-new--20250508"></a>
+
+<a id="whats-new--may-8u002c-2025"></a>
+
+#### May 8, 2025
+
+| New guidance or change | Page |
+| --- | --- |
+| Removed outdated language that indicated that a list of options is treated differently than other unordered lists. | [Lists](#lists) |
+| Added a page about writing prescriptive documentation. | [Prescriptive documentation](#prescriptive-documentation), [Word list](#word-list--should) |
+| Removed guidance that said to include empty parentheses after method names. | [Code in text](#code-in-text) |
+| Changed footnotes guidance to recommend using numbers instead of symbols. | [Footnotes](#footnotes) |
+| Simplified contractions guidance, removing excess explanations and examples. | [Contractions](#contractions) |
+| Added examples to guidance about introducing sections of a document. | [Headings and titles](#headings) |
+| Added an example service account ID. | [Example domains and names](#examples) |
+| Simplified and clarified guidance about articles, including using articles before abbreviations and product names. | [Abbreviations](#abbreviations), [Articles (a, an, the)](#articles), [Product names](#product-names), [Word list](#word-list) |
+| Included a general explanation of why we don't document future features. | [Document future features](#future) |
+| Revised guidance to recommend using an abbreviation in a title or heading only if the abbreviation is the more commonly known version of the word. | [Headings and titles](#headings) |
+| Added a suggestion that when you use an imperative in running text, consider whether to write a procedure instead. | [Second person and first person](#person) |
+| Generalized existing guidance on what to do when jargon is part of a command or code sample. | [Jargon](#jargon) |
+| Consolidated guidance about cross-references and linking into one page. | [Cross-references and linking](#cross-references) |
+| Added some specific reasons why we avoid directional language when orienting the reader to information on a page. | [Write accessible documentation](#accessibility) |
+| Clarified that spacing around icons is a judgment call based on readability. | [UI elements and interaction](#ui-elements) |
+| Removed an example phone number. | [Example domains and names](#examples) |
+| Removed the word list entry for *property* because its usage heavily depends on the technical context. | [Word list](#word-list) |
+ New word list entries: *choose*, *confidential*, *sensitive*, *image*, *Fast Healthcare Interoperability Resources (FHIR)* | [Word list](#word-list) |
+
+<a id="whats-new--20250117"></a>
+
+<a id="whats-new--january-17u002c-2025"></a>
+
+#### January 17, 2025
+
+| New guidance or change | Page |
+| --- | --- |
+| Consolidated spelling guidance into the introduction to the word list. Extended guidance about how to use the preferred dictionary to determine which spelling to use for a word with multiple spellings. | [Word list](#word-list) |
+| Generalized guidance about when to use spaces or tabs for indentation in code samples. | [Code samples](#code-samples--guidelines) |
+| Changed guidance for formatting telephone numbers to recommend using dashes instead of parentheses to set the area code off from the rest of the number. | [Format phone numbers](#phone-numbers) |
+| Extended and clarified recommendation to avoid the abbreviations *i.e.*, *e.g.*, and *etc.* in most cases. | [Abbreviations](#abbreviations--dont-use), [Comma-separated lists](#lists--comma-separated-lists) |
+
+<a id="whats-new--20241029"></a>
+
+<a id="whats-new--october-29u002c-2024"></a>
+
+#### October 29, 2024
+
+| New guidance or change | Page |
+| --- | --- |
+| Consolidated and expanded guidance about hyphens and closed compounds. | [Hyphens](#hyphens) |
+| Added *hotspot* to the word list. | [Word list](#word-list--hotspot) |
+
+<a id="whats-new--20240815"></a>
+
+<a id="whats-new--august-15u002c-2024"></a>
+
+#### August 15, 2024
+
+| New guidance or change | Page |
+| --- | --- |
+| Added guidance for distinguishing between binary and decimal units, such as gibibytes (GiB) and gigabytes (GB). Also corrected abbreviation of *kilobyte* to *kB*. | [Decimal and binary units](#units-of-measure--byte-units) |
+| Expanded guidance for referring to figures and other images in text. Clarified that figure numbers are not required. | [Figure captions](#images--figure-captions) |
+| Added word list entries: *curl*, *whitepaper*, *long-running operation* | [Word list](#word-list) |
+| Updated word list entries: *ingest*, *execute*, *content type*, *media type*, *MIME* | [Word list](#word-list) |
+
+<a id="whats-new--20240516"></a>
+
+<a id="whats-new--may-16u002c-2024"></a>
+
+#### May 16, 2024
+
+| New guidance or change | Page |
+| --- | --- |
+| Added the term *generative AI*. | [Word list](#word-list--generative-ai) |
+| Added the term *rehost*. Redirected the definition for *lift and shift* to *rehost*. | [Word list](#word-list--rehost) |
+| Clarified that the style guide shows examples of how placeholders render but doesn't explain how to implement this visual styling. | [Format placeholders](#placeholders) |
+| Added an example to the guidance showing how to use site-root-relative URLs to link to another page on the same server. | [Cross-references](https://developers.google.com/style/cross-references#same-server) |
+| Clarified guidance about how to format anchor text. | [Make headings into link targets](#headings-targets--markdown) |
+
+<a id="whats-new--20240321"></a>
+
+<a id="whats-new--march-21u002c-2024"></a>
+
+#### March 21, 2024
+
+| New guidance or change | Page |
+| --- | --- |
+| Expanded guidance about avoidance of the term *drop-down*. | [Word list](#word-list--drop-down) |
+| Added examples to guidance about writing documentation that focuses on the present state of the software. | [Timeless documentation](#timeless-documentation) |
+| Added recommendation to use a more precise term than *workload* when possible, or to define what the term means in the specific context.  Added related guidance about avoiding ambiguous or overloaded words like *workload*, *solution*, and *support*, or defining them in each context. | [Word list](#word-list--workload),  [Jargon](#jargon) |
+| Aligned guidance with XML and HTML specifications to recommend against the use of angle brackets as part of an element name, but instead to only use angle brackets as part of a tag. | [Code in text](#code-in-text) |
+
+<a id="whats-new--20240122"></a>
+
+<a id="whats-new--january-22u002c-2024"></a>
+
+#### January 22, 2024
+
+| New guidance or change | Page |
+| --- | --- |
+| Cleaned up word list by removing entries that only provided [hyphenation](#hyphens), [spelling](https://developers.google.com/style/spelling), or [abbreviation](#abbreviations) guidance that followed directly from our general guidance. | [Word list](https://developers.google.com/style/style/word-list) |
+| Consolidated and clarified guidance regarding alt text, figure captions, and figure descriptions. | [Diagrams, figures, and other images](#images)  [Write accessible documentation](#accessibility) |
+| Added recommendation to avoid linking to other document sets from navigation controls such as a table of contents. | [Links to other sites](https://developers.google.com/style/links-external) |
+| Clarified guidance about avoiding words such as *above* and *below* in references to documentation and user interfaces for accessibility reasons, and provided example of appropriate non-directional usage. | [Word list](#word-list--above) |
+| Removed prohibition against hyphenation of the phrase *open source*, so this term now follows our general guidance for hyphenation, which allows for hyphenation of an adjectival phrase to add clarity and remove ambiguity. | [Word list](https://developers.google.com/style/word-list#open-source) |
+| Changed guidance for indicating the omission of code from an instructional code snippet to recommend the use of an explanatory comment instead of a mere ellipsis. | [Code samples](#code-samples) |
+
+<a id="whats-new--20231108"></a>
+
+<a id="whats-new--november-8u002c-2023"></a>
+
+#### November 8, 2023
+
+| New guidance or change | Page |
+| --- | --- |
+| Changed guidance regarding run-in headings in description lists to recommend that the punctuation (such as a colon) is not formatted as bold. Making the punctuation bold caused the punctuation to seem to be part of the heading string, which caused confusion in cases such as UI labels. | [Lists](#lists--description-lists-that-use-run-in-headings) |
+| Removed the recommendation to use a special *external* icon (indicated by `class="external"`) for links. Readers and writers have expressed confusion about the meaning and usage of this icon. We strengthened guidance about using other, explicit means to inform the reader about the destination and behavior of a link. | [Link text](https://developers.google.com/style/link-text#write-link-text) |
+
+<a id="whats-new--20231024"></a>
+
+<a id="whats-new--october-24u002c-2023"></a>
+
+#### October 24, 2023
+
+| New guidance or change | Page |
+| --- | --- |
+| Removed page about custom font styling, which only said to use styles defined in the style sheet for the website. Redirected link to page about HTML formatting and semantic tagging. | [HTML and semantic tagging](https://developers.google.com/style/fonts) |
+| Added word-list entry for *then* and expanded entry for *if* to recommend the use of the optional helper word *then* in many cases in which it might be omitted in casual usage. | [Word list](#word-list--then) |
+| Simplified and unified guidance for *jank* and *janky* to recommend that these terms only be used for specific graphics issues. | [Word list](#word-list--jank) |
+| Strengthened guidance against the use of *and/or* except in cases where space is limited. | [Slashes](#slashes--and-or) |
+
+<a id="whats-new--20230929"></a>
+
+<a id="whats-new--september-29u002c-2023"></a>
+
+#### September 29, 2023
+
+| New guidance or change | Page |
+| --- | --- |
+| Added several examples of when and how to use quotation marks. | [Quotation marks](#quotation-marks) |
+| Added explanation of guidance against anthropomorphism. | [Anthropomorphism](#anthropomorphism) |
+| Added guidance about using a hyphen with the prefix *non* before hyphenated compounds. | [Hyphens](#hyphens) |
+
+<a id="whats-new--20230824"></a>
+
+<a id="whats-new--august-24u002c-2023"></a>
+
+#### August 24, 2023
+
+| New guidance or change | Page |
+| --- | --- |
+| Added link buttons to each entry in the word list to make deep-linking to individual entries easier. | [Word list](#word-list) |
+| Extended guidance about example names to recommend using an initial to represent a person's surname. | [Example person surnames](#examples--example-person-surnames) |
+| Clarified guidance about when to use present tense and when to use future tense. | [Present tense](#tense) |
+| Extended link text guidance to include an example for `mailto` links. | [Link text](https://developers.google.com/style/link-text) |
+
+<a id="whats-new--20230726"></a>
+
+<a id="whats-new--july-26u002c-2023"></a>
+
+#### July 26, 2023
+
+| New guidance or change | Page |
+| --- | --- |
+| Added link to [Google API guidelines](https://google.aip.dev/192) for information about code comments. | [API reference code comments](#api-reference-comments) |
+| Revised guidance for the word *toggle* to recommend against use as a verb. | [UI elements and interaction](#ui-elements--term-toggle) |
+| Extended and clarified guidance for names for directories and files. | [Filenames and file types](#filenames) |
+| Added explanation for why to use code format for code items. | [Code in text](#code-in-text) |
+
+<a id="whats-new--20230615"></a>
+
+<a id="whats-new--june-15u002c-2023"></a>
+
+#### June 15, 2023
+
+| New guidance or change | Page |
+| --- | --- |
+| Consolidated guidance about periods and end punctuation. Also removed standalone pages about exclamation points and about spacing after periods. | [Periods and other end punctuation](#periods) |
+| Revised guidance about hyphens to suggest a lookup strategy, categorize uses, and note exceptions. | [Hyphens](#hyphens) |
+| Created firmer guidance about punctuation in lists for run-in headings and at the end of list items. | [Lists](#lists) |
+| Strengthened capitalization guidance: when *not* to use capitalization, and how to use capitalization with product names. | [Capitalization](#capitalization), [Product names](#product-names) |
+| Added guidance about using end punctuation when documenting a command-line option or argument. | [Document command-line syntax](#code-syntax) |
+| Improved description and examples for using first-person pronouns (*we*, *our*). | [Second person and first person](#person) |
+| Updated word list entries: *etc.*, *OK*, *user*, *we*, *you* | [Word list](#word-list) |
+
+<a id="whats-new--20230509"></a>
+
+<a id="whats-new--may-9u002c-2023"></a>
+
+#### May 9, 2023
+
+| New guidance or change | Page |
+| --- | --- |
+| Softened guidance regarding the choice between the pronouns *who* and *that*. | [Pronouns](#pronouns) |
+| Added *web interface* as an alternative to *console* and *UI* in general references to a browser-based interface. | [Word list](#word-list--ui) |
+| Clarified the purpose of the page about phone number formats. | [Format phone numbers in text](#phone-numbers) |
+
+<a id="whats-new--20230331"></a>
+
+<a id="whats-new--march-31u002c-2023"></a>
+
+#### March 31, 2023
+
+| New guidance or change | Page |
+| --- | --- |
+| Expanded and clarified guidance about using trademarks only as modifiers. This guidance emphasizes that you should never modify a trademark, such as by creating a possessive or plural form. | [Use trademarks only as modifiers](#trademarks--use-trademarks-only-as-modifiers) |
+| Strengthened guidance against shortening product names to anything other than an approved alternative name. | [Google product names](#product-names--shortening) |
+| Expanded guidance about using ARIA labels in text that describes icons in graphical user interfaces. This improves accessibility and increases consistency in terminology used to refer to visual elements in text. | [Buttons and icons](#ui-elements--buttons), [Accessibility](#accessibility) |
+| New word list entry: *toolkit* | [Word list](#word-list) |
+
+<a id="whats-new--20230213"></a>
+
+<a id="whats-new--february-13u002c-2023"></a>
+
+#### February 13, 2023
+
+| New guidance or change | Page |
+| --- | --- |
+| Added a page about paragraph structure, which provides guidance about recommended paragraph length and order of information. | [Paragraph structure](#paragraph-structure) |
+| Expanded accessibility guidance to say that a document should convey its information when you use it without images or animation. | [Write accessible documentation](#accessibility) |
+| Clarified that letter keys should be represented with uppercase letters. | [Press and type keyboard keys](#ui-elements--keyboard) |
+| Added *existing* to list of examples of potentially problematic words in timeless documentation. | [Timeless documentation](#timeless-documentation) |
+| Recommended using an empty `alt` attribute for icons that include a text label. | [Buttons and icons](#ui-elements--buttons) |
+| Strengthened and clarified guidance about avoiding culturally specific references and about using simple and consistent language. | [Voice and tone](#tone) |
+| Created section about items that are sometimes—but not always—formatted in code font, such as email addresses. | [Items that are sometimes in code font](#code-in-text--sometimes) |
+| Added information about example internationalized domain names. | [Example domain names](#examples--domains) |
+| Expanded guidance about choosing example email addresses. | [Example email addresses](#examples--addresses) |
+| Expanded guidance about using second-person *you* to refer to the reader of a document and, generally, using third-person *user* to refer to the intended user of the software that the reader is developing. | [Second person](#person) |
+
+<a id="whats-new--20221212"></a>
+
+<a id="whats-new--december-12u002c-2022"></a>
+
+#### December 12, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| New word list entry: *anti-pattern* | [Word list](#word-list) |
+
+<a id="whats-new--20221107"></a>
+
+<a id="whats-new--november-7u002c-2022"></a>
+
+#### November 7, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| To emphasize a negative, use `<em>`not`</em>`. | [Contractions](#contractions) |
+
+<a id="whats-new--20221031"></a>
+
+<a id="whats-new--october-31u002c-2022"></a>
+
+#### October 31, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| New word list entry: *standalone* | [Word list](#word-list) |
+
+<a id="whats-new--20221003"></a>
+
+<a id="whats-new--october-10u002c-2022"></a>
+
+#### October 10, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Added guidance about how to document optional arguments for commands.  Special characters that indicate optional and mutually exclusive arguments in commands—such as brackets, braces, and pipes—break commands if the user doesn't edit them first. The new guidance offers several approaches for avoiding these problems. | [Code in text](#code-in-text), [Code samples](#code-samples), [Document command-line syntax](#code-syntax) |
+| Clarified that contractions are recommended in general, but not required in all cases. | [Contractions](#contractions) |
+
+<a id="whats-new--20220926"></a>
+
+<a id="whats-new--september-26u002c-2022"></a>
+
+#### September 26, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| In figure captions, always use end punctuation, and use complete sentences when possible. | [Figures and other images](#images) |
+| Added separators between word list terms, to improve readability. | [Word list](#word-list) |
+
+<a id="whats-new--20220919"></a>
+
+<a id="whats-new--september-19u002c-2022"></a>
+
+#### September 19, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Added guidance to multiple pages about ways to make documentation more inclusive for readers who have a variety of cognitive patterns. | [Write accessible documentation](#accessibility), [Cross-references](#cross-references), [Numbers](#numbers), [Procedures](#procedures) |
+
+<a id="whats-new--20220912"></a>
+
+<a id="whats-new--september-12u002c-2022"></a>
+
+#### September 12, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Don't present new information in tables through images or symbols alone. | [Tables](#tables) |
+| Clarified guidance about using footnotes in tables. | [Tables](#tables) |
+| Added instructions for how to look up a UI element's `aria-label` attribute. | [UI elements and interaction](#ui-elements) |
+
+<a id="whats-new--20220905"></a>
+
+<a id="whats-new--september-5u002c-2022"></a>
+
+#### September 5, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Strengthened recommendation to avoid using semicolons where possible, and removed basic information about semicolons.  Our [accessibility guidance](#accessibility) recommends against using semicolons where possible, because screen readers may not clearly indicate them. So we changed our semicolon guidance to be more in line with our accessibility guidance. | [Semicolons](#semicolons) |
+| Expanded and clarified guidance about what to put in code font. | [Code in text](#code-in-text) |
+
+<a id="whats-new--20220829"></a>
+
+<a id="whats-new--august-29u002c-2022"></a>
+
+#### August 29, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Changed and clarified recommended phrasing for describing boolean parameters in reference docs. | [API reference code comments](#api-reference-comments) |
+| Expanded and clarified explanation of why we use straight quotation marks and apostrophes. | [Quotation marks](#quotation-marks) |
+| Added suggested alternative terms for *cloud-native*. | [Word list](#word-list) |
+| New word list entries: *prebuilt*, *scroll* | [Word list](#word-list) |
+
+<a id="whats-new--20220822"></a>
+
+<a id="whats-new--august-22u002c-2022"></a>
+
+#### August 22, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Clarified that *and then* is generally better than just *then*. | [Write for a global audience](#translation) |
+| New word list entries: *canary*, *multi-service*, *precapture*, *pre-existing*, *presubmit*, *rebranding*, *roll out* | [Word list](#word-list) |
+
+<a id="whats-new--20220815"></a>
+
+<a id="whats-new--august-15u002c-2022"></a>
+
+#### August 15, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Expanded table-formatting guidance, to improve accessibility. | [Tables](#tables) |
+| Improved guidance about figure captions, descriptions, and alt text. | [Figures and other images](#images) |
+
+<a id="whats-new--20220808"></a>
+
+<a id="whats-new--august-8u002c-2022"></a>
+
+#### August 8, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Write *a SQL* rather than *an SQL*.  Both are in use, but *a SQL* is significantly more common. | [Articles (a, an, the)](#articles) |
+| Added information about our distinction between *don't use* and *avoid*. | [Word list](#word-list) |
+| New word list entries: *nonce*, *SQL* | [Word list](#word-list) |
+
+<a id="whats-new--20220801"></a>
+
+<a id="whats-new--august-1u002c-2022"></a>
+
+#### August 1, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Expanded guidance about *mobile* and related terms. | [Word list](#word-list) |
+| It's OK to use *below* in set phrases such as *below (the) average*. | [Word list](#word-list) |
+| Expanded our information about serial commas. | [Commas](#commas) |
+| To refer to a file with the `.tiff` extension, use the phrase *TIFF file*. | [Filenames and file types](#filenames) |
+| New word list entries: *admin*; *blue-green*; *cold*, *hot*, and *warm* (in the context of a failover, spare, or standby); *inline*; *mobile phone*; *online* | [Word list](#word-list) |
+
+<a id="whats-new--20220725"></a>
+
+<a id="whats-new--july-25u002c-2022"></a>
+
+#### July 25, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Updated guidance about the term *Cloud console*. | [Word list](#word-list) |
+| Clarified guidance about the term *see*. | [Word list](#word-list) |
+
+<a id="whats-new--20220718"></a>
+
+<a id="whats-new--july-18u002c-2022"></a>
+
+#### July 18, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Revised guidance about using the term *element* in HTML and XML contexts. | [Word list](#word-list) |
+| Revised and expanded guidance about how to form possessives. | [Possessives](#possessives) |
+| New word list entry: *tag* | [Word list](#word-list) |
+
+<a id="whats-new--20220620"></a>
+
+<a id="whats-new--june-20u002c-2022"></a>
+
+#### June 20, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| New word list entry: *brown bag* | [Word list](#word-list) |
+
+<a id="whats-new--20220613"></a>
+
+<a id="whats-new--june-13u002c-2022"></a>
+
+#### June 13, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Clarified guidance about placement of *only*. | [Write for a global audience](#translation) |
+| New word list entry: *using* | [Word list](#word-list) |
+
+<a id="whats-new--20220606"></a>
+
+<a id="whats-new--june-6u002c-2022"></a>
+
+#### June 6, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Clarified guidance about *Interconnect connection*. | [Word list](#word-list) |
+
+<a id="whats-new--20220516"></a>
+
+<a id="whats-new--may-16u002c-2022"></a>
+
+#### May 16, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Offset footnote symbols using superscript. | [Footnotes](#footnotes) |
+| New word list entries: *could*, *would* | [Word list](#word-list) |
+
+<a id="whats-new--20220509"></a>
+
+<a id="whats-new--may-9u002c-2022"></a>
+
+#### May 9, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Strengthened guidance recommending avoiding humor in documentation.  Most humor is difficult to translate, and much humor is culturally specific. | [Write for a global audience](#translation) |
+| Changed the link to a resource about identity-first language.  The site that we had previously linked to has disappeared. | [Write inclusive documentation](#inclusive-documentation) |
+
+<a id="whats-new--20220425"></a>
+
+<a id="whats-new--april-25u002c-2022"></a>
+
+#### April 25, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Added a new page about jargon. | [Jargon](#jargon) |
+| Expanded guidance about when to use the various notice types. | [Notes, cautions, warnings, and other notices](#notices) |
+| Clarified and expanded guidance about when to remove locales from URLs. | [Link to other sites](https://developers.google.com/style/links-external) |
+| New word list entries: *final solution*, `gsutil` | [Word list](#word-list) |
+
+<a id="whats-new--20220418"></a>
+
+<a id="whats-new--april-18u002c-2022"></a>
+
+#### April 18, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Introduce an interactive element (such as a button that expands and collapses) in the text preceding the element, to improve accessibility. | [Write accessible documentation](#accessibility) |
+
+<a id="whats-new--20220411"></a>
+
+<a id="whats-new--april-11u002c-2022"></a>
+
+#### April 11, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Changed guidance about taking screenshots. | [Figures and other images](#images) |
+| In code samples, indicate omitted code using three dots and no spaces (`...`) | [Code samples](#code-samples) |
+| New word list entries: *+* (appended to numbers in text) | [Word list](#word-list) |
+
+<a id="whats-new--20220404"></a>
+
+<a id="whats-new--april-4u002c-2022"></a>
+
+#### April 4, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Removed guidance about using lettered lists for mutually exclusive options.  The semantic distinction that we were making isn't in wide use, and lettered lists aren't supported in standard Markdown, so we no longer recommend using lettered lists. | [Lists](#lists) |
+| Clarified recommendation about how to italicize in Markdown. | [Text-formatting summary](#text-formatting) |
+| Changed guidance about *dead-letter queue* and *hold the pointer over*. | [Word list](#word-list) |
+| New word list entry: *break-glass* | [Word list](#word-list--break-glass) |
+
+<a id="whats-new--20220328"></a>
+
+<a id="whats-new--march-28u002c-2022"></a>
+
+#### March 28, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Clarified guidance about using an introductory phrase before the output of a command. | [Document command-line syntax](#code-syntax) |
+| New word list entry: *Unicode* | [Word list](#word-list--unicode) |
+
+<a id="whats-new--20220221"></a>
+
+<a id="whats-new--february-21u002c-2022"></a>
+
+#### February 21, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Use `gcloud` *CLI* instead of `gcloud` *command-line tool*. | [Word list](#word-list) |
+| New word list entry: *shift left* | [Word list](#word-list) |
+
+<a id="whats-new--20220207"></a>
+
+<a id="whats-new--february-7u002c-2022"></a>
+
+#### February 7, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| In general, don't use a single *x* or a series of *x*'s as placeholders; instead, use a more informative placeholder. | [Formatting placeholders](#placeholders) |
+| New word list entry: *GBps* | [Word list](#word-list) |
+
+<a id="whats-new--20220131"></a>
+
+<a id="whats-new--january-31u002c-2022"></a>
+
+#### January 31, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Avoid repeating the exact page title as a heading on the page. | [Headings and titles](#headings) |
+| Expanded guidance about *runtime* and *run time*. | [Word list](#word-list) |
+| Added *.adoc* and *.md* to the list of examples of filename extensions. | [Filenames and file types](#filenames--file-type-names) |
+
+<a id="whats-new--20220124"></a>
+
+<a id="whats-new--january-24u002c-2022"></a>
+
+#### January 24, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| Clarified guidance about video formats.  The main reason to avoid using animated GIF is that it's resource-inefficient. | [Figures and other images](#images) |
+| Clarified guidance about changing an existing custom anchor for a heading. | [Making headings into link targets](#headings-targets) |
+
+<a id="whats-new--20220118"></a>
+
+<a id="whats-new--january-18u002c-2022"></a>
+
+#### January 18, 2022
+
+| New guidance or change | Page |
+| --- | --- |
+| For animations and videos, use a compressed format (such as MP4), not animated GIF. | [Figures and other images](#images) |
+| Don't use *email* as a verb. | [Word list](#word-list) |
+| New word list entry: *healthcare* | [Word list](#word-list) |
+
+<a id="whats-new--20211231"></a>
+
+<a id="whats-new--december-31u002c-2021"></a>
+
+#### December 31, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Added headings to make guidance about using simple and unambiguous language easier to find. | [Writing for a global audience](https://developers.google.com/style/translation#clear-language) |
+| Expanded guidance about demonstrative pronouns. | [Ambiguous pronoun references](#pronouns--ambiguous-pronoun-references) |
+| Clarified guidance about when to use code format for URLs. | [Code in text](#code-in-text) |
+| Refreshed guidance about optimizing for search (SEO). | [Text for images](#images--text-associated-with-images),  [Link text](https://developers.google.com/style/link-text),  [Filenames](#filenames),  [Headings](#headings) |
+
+<a id="whats-new--20211130"></a>
+
+<a id="whats-new--november-30u002c-2021"></a>
+
+#### November 30, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Added examples and clarified guidance recommending use of optional pronouns and other helper words. | [Optional pronouns](#pronouns--optional-pronouns),  [Writing for a global audience](https://developers.google.com/style/translation#write-short,-clear,-and-precise-sentences) |
+| Added guidance for linking to a section of another page. | [Links to sections on another page](https://developers.google.com/style/cross-references#different-page) |
+| New word list entries: *portal*, *copy and paste*, *performant* | [Word list](#word-list) |
+
+<a id="whats-new--20211101"></a>
+
+<a id="whats-new--november-1u002c-2021"></a>
+
+#### November 1, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Added guidance about expanding shortened words and symbols. | [Abbreviations](#abbreviations--dont-use) |
+| Expanded guidance about the order of information in complex procedures, including the recommended order for goals, results, and justifications for steps. | [Procedures](https://developers.google.com/style/procedures#order-of-information-in-complex-steps) |
+| Added guidance about including a description of default behaviors in API reference documents. | [API reference code comments](#api-reference-comments--parameters) |
+| Reinforced guidance about including words like *file* after items in code font, such as filenames. | [Filenames](#filenames) |
+| Added guidance about using alternatives to directional terms such as *above* and *below* when referring to location in a document. | [Writing accessible documentation](#accessibility--document-rendering) |
+| New word list entries: *practitioner*, *left-nav, right-nav*, *create new* | [Word list](#word-list) |
+
+<a id="whats-new--20211011"></a>
+
+<a id="whats-new--october-11u002c-2021"></a>
+
+#### October 11, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Simplified placeholder examples and clarified when to repeat an explanation for a placeholder. | [Formatting placeholders](https://developers.google.com/style/placeholders#explaining-placeholders) |
+| Added guidance on using brackets and dots (`[...]`) to indicate omitted output in an example. | [Output from commands](#code-syntax--output) |
+
+<a id="whats-new--20211004"></a>
+
+<a id="whats-new--october-4u002c-2021"></a>
+
+#### October 4, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Added page about philosophy and principles of the style guide. | [Philosophy of this style guide](#philosophy) |
+| Clarified recommendation to include a noun after a code element and to avoid inflecting code elements. | [Code in text](#code-in-text) |
+| Added guidance to not use numbers in headings to indicate a sequence of sections. | [Headings and titles](#headings) |
+| Updated examples for verb forms in headings and titles, including special cases such as *Pricing* and *Billing*. | [Headings and titles](#headings) |
+
+<a id="whats-new--20210913"></a>
+
+<a id="whats-new--september-13u002c-2021"></a>
+
+#### September 13, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Clarified guidance about specifying the context for a task. | [Procedures](#procedures) |
+| Avoid using *run the following command* to introduce code. Instead, focus on what the command does. | [Procedures](#procedures) |
+| Expanded the guidance for *CLI*. | [Word list](#word-list--cli) |
+| New word list entries: *ransomware* | [Word list](#word-list) |
+
+<a id="whats-new--20210906"></a>
+
+<a id="whats-new--september-6u002c-2021"></a>
+
+#### September 6, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Avoid words and phrases that anchor the documentation to a point in time or assume knowledge of prior or future products and features. | [Timeless documentation](#timeless-documentation) |
+| Added *navigation menu* to the list of UI terms. | [UI elements and interaction](#ui-elements) |
+| Marked *slice and dice* as *don't use*. | [Word list](#word-list--slice) |
+| New word list entries: *web application firewall*, *while* | [Word list](#word-list) |
+
+<a id="whats-new--20210830"></a>
+
+<a id="whats-new--august-30u002c-2021"></a>
+
+#### August 30, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Introduce a list of placeholders with the phrase *This output includes the following values:*  The old guidance for what introductory phrase to use wasn't a complete sentence. | [Formatting placeholders](#placeholders) |
+
+<a id="whats-new--20210816"></a>
+
+<a id="whats-new--august-16u002c-2021"></a>
+
+#### August 16, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Added guidance for heading and title phrasing.  Many writers ask what verb forms to use in headings and titles. The new guidance helps writers make that decision more easily, and it improves consistency for translators and readers. | [Headings and titles](#headings) |
+
+<a id="whats-new--20210802"></a>
+
+<a id="whats-new--august-2u002c-2021"></a>
+
+#### August 2, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| When you need an example project name, create a name that's meaningful or descriptive. | [Example domains and names](#examples) |
+| New word list entry: *pop-up* | [Word list](#word-list--pop-up) |
+
+<a id="whats-new--20210726"></a>
+
+<a id="whats-new--july-26u002c-2021"></a>
+
+#### July 26, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| New word list entry: *intercluster* | [Word list](#word-list--intercluster) |
+
+<a id="whats-new--20210719"></a>
+
+<a id="whats-new--july-19u002c-2021"></a>
+
+#### July 19, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Added guidance about whether to include an abbreviation if it is used only once in the document. | [Abbreviations](#abbreviations) |
+| Clarified and added guidance about the terms *above*, *below*, *earlier*, *higher*, *later*, *lower*, and *under* | [Word list](#word-list) |
+
+<a id="whats-new--20210712"></a>
+
+<a id="whats-new--july-12u002c-2021"></a>
+
+#### July 12, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Don't refer to casing styles by names like *camel case* or *snake case*. | [Capitalization](#capitalization--special-capitalization-style-names) |
+
+<a id="whats-new--20210705"></a>
+
+<a id="whats-new--july-5u002c-2021"></a>
+
+#### July 5, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| In the first sentence of each procedural step, include an imperative verb. | [Procedures](#procedures) |
+
+<a id="whats-new--20210628"></a>
+
+<a id="whats-new--june-28u002c-2021"></a>
+
+#### June 28, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| New word list entries: *can*, *might*, *must*  A writer asked us to clarify our guidance for these terms. | [Word list](#word-list) |
+
+<a id="whats-new--20210621"></a>
+
+<a id="whats-new--june-21u002c-2021"></a>
+
+#### June 21, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Updated guidance on fictitious person names, and added a list of specific given names to use in examples.  Many writers have asked for a list of recommended names to use. | [Example domains and names](#examples--names) |
+
+<a id="whats-new--20210607"></a>
+
+<a id="whats-new--june-7u002c-2021"></a>
+
+#### June 7, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Don't use UI element labels as if they were English verbs or nouns. | [UI elements and interaction](#ui-elements--referring-to-ui) |
+
+<a id="whats-new--20210524"></a>
+
+<a id="whats-new--may-24u002c-2021"></a>
+
+#### May 24, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Clarified phrasing of guidance about avoiding using the same link text for different target pages. | [Link text](https://developers.google.com/style/link-text) |
+| Added guidance about what to do when link text includes a command or other code-font element. | [Link text](https://developers.google.com/style/link-text) |
+| Try to keep the main subject and verb as close to the beginning of a sentence as possible. | [Writing for a global audience](#translation) |
+| Updated guidance about using a hyphen to indicate a range of numbers.  For example, we now recommend using an ordinary hyphen rather than a nonbreaking hyphen, for ease of authoring. | [Hyphens](#hyphens) |
+| Removed the word list entry for *AJAX*.  The entry was outdated, and Wikipedia covers the term better. | [Word list](#word-list) |
+| New word list entries: *webmaster*, *white label*  Use more specific terms. | [Word list](#word-list) |
+
+<a id="whats-new--20210503"></a>
+
+<a id="whats-new--may-3u002c-2021"></a>
+
+#### May 3, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Don't put a link in a heading.  A link in a heading is easy for a reader to miss. | [Headings and titles](https://developers.google.com/style/headings#formatting-a-heading-or-title) |
+| Clarified and reorganized the material about which preposition to use with each UI element. | [UI elements and interaction](#ui-elements--prepositions) |
+
+<a id="whats-new--20210426"></a>
+
+<a id="whats-new--april-26u002c-2021"></a>
+
+#### April 26, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Clarified guidance about when to create a custom anchor for a heading. | [Making headings into link targets](#headings-targets) |
+
+<a id="whats-new--20210419"></a>
+
+<a id="whats-new--april-19u002c-2021"></a>
+
+#### April 19, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Write one-step procedures with a bullet. | [Procedures](#procedures) |
+| In headings, use punctuation sparingly. | [Headings and titles](#headings) |
+| Added a link to another resource about active and passive voice. | [Active voice](#voice) |
+| Clarified guidance about the word *native*. | [Word list](#word-list) |
+| New entry: *sherpa* | [Word list](#word-list) |
+
+<a id="whats-new--20210412"></a>
+
+<a id="whats-new--april-12u002c-2021"></a>
+
+#### April 12, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Clarified that guidance that applies to text generally also applies to headings. | [Headings and titles](#headings) |
+| Strengthened recommendation to use contractions. | [Contractions](#contractions) |
+| New entry: *downscope* | [Word list](#word-list) |
+
+<a id="whats-new--20210405"></a>
+
+<a id="whats-new--april-5u002c-2021"></a>
+
+#### April 5, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Expanded guidance about using simple words and phrases. | [Writing for a global audience](#translation) |
+| New entry: *instance group* | [Word list](#word-list) |
+
+<a id="whats-new--20210329"></a>
+
+<a id="whats-new--march-29u002c-2021"></a>
+
+#### March 29, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| To form the plural of an abbreviation, treat the abbreviation like an English word, and follow standard English pluralizing patterns. | [Abbreviations](#abbreviations) |
+
+<a id="whats-new--20210315"></a>
+
+<a id="whats-new--march-15u002c-2021"></a>
+
+#### March 15, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| New entries: more than 60 terms to not use or to use with caution, for inclusive-language reasons (too many to list here) | [Word list](#word-list) |
+
+<a id="whats-new--20210308"></a>
+
+<a id="whats-new--march-8u002c-2021"></a>
+
+#### March 8, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Expanded and clarified guidance about placeholders. | [Formatting placeholders](#placeholders) |
+| New entries: *outside the box* (and related terms), and more than 40 networking terms (too many to list here) | [Word list](#word-list) |
+
+<a id="whats-new--20210301"></a>
+
+<a id="whats-new--march-1u002c-2021"></a>
+
+#### March 1, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Clarified guidance about spelling out an abbreviation. | [Abbreviations](#abbreviations) |
+| Use *documentation set* instead of *docset*. | [Word list](#word-list) |
+
+<a id="whats-new--20210215"></a>
+
+<a id="whats-new--february-15u002c-2021"></a>
+
+#### February 15, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| New entry: *lock-in* | [Word list](#word-list) |
+
+<a id="whats-new--20210201"></a>
+
+<a id="whats-new--february-1u002c-2021"></a>
+
+#### February 1, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Revised and clarified guidance about using *should*. | [Word list](#word-list--should) |
+| In some contexts, you can use *active*/*standby* as a replacement for *master*/*slave*. | [Word list](#word-list--slave) |
+| Updated guidance about using *firewalls*. | [Word list](#word-list--firewalls) |
+| New entries: *black-box*, *external IP address*, *gray-box*, *IPsec*, *N/A*, *war room*, *white-box* | [Word list](#word-list) |
+
+<a id="whats-new--20210125"></a>
+
+<a id="whats-new--january-25u002c-2021"></a>
+
+#### January 25, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Updated guidance about link text to remove reliance on external sources. | [Link text](https://developers.google.com/style/link-text) |
+| Don't use abbreviations or product names as verbs. | [Abbreviations](#abbreviations--verbs), [Product names](#product-names--product-names-as-verbs) |
+
+<a id="whats-new--20210111"></a>
+
+<a id="whats-new--january-11u002c-2021"></a>
+
+#### January 11, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Clarified guidance about indentation in code samples. | [Code samples](#code-samples) |
+| Added links to more coding-style guides. | [Code in text](#code-in-text) |
+| Clarified that you can use *please* under certain specific circumstances. | [Word list](#word-list--please) |
+| Clarified guidance about *terminate*. | [Word list](#word-list--terminate) |
+| New entries: *hang*, *MTU*, *utilize* | [Word list](#word-list) |
+
+<a id="whats-new--20210104"></a>
+
+<a id="whats-new--january-4u002c-2021"></a>
+
+#### January 4, 2021
+
+| New guidance or change | Page |
+| --- | --- |
+| Merged landing page with "Other editorial resources" to help new readers learn how to use this guide. | [About this guide](#style) |
+| New entry: *deprecate* | [Word list](#word-list) |
+
+<a id="whats-new--20201228"></a>
+
+<a id="whats-new--december-28u002c-2020"></a>
+
+#### December 28, 2020
+
+| New guidance or change | Page |
+| --- | --- |
+| Added an explanation of what placeholders are. | [Formatting placeholders](#placeholders) |
+
+<a id="whats-new--20201214"></a>
+
+<a id="whats-new--december-14u002c-2020"></a>
+
+#### December 14, 2020
+
+| New guidance or change | Page |
+| --- | --- |
+| Distinguished between the Google Cloud product called Identity and Access Management (IAM) and the general practice of identity and access management. | [Word list](#word-list--iam) |
+| Clarified that the word *terminate* has a specific meaning in telephony. | [Word list](#word-list--terminate) |
+
+<a id="whats-new--20201207"></a>
+
+<a id="whats-new--december-7u002c-2020"></a>
+
+#### December 7, 2020
+
+| New guidance or change | Page |
+| --- | --- |
+| When you describe navigating through menus using angle brackets, add an `aria-label` attribute to each angle bracket, for use by screen readers: `<span aria-label="and then">></span>`. | [UI elements and interaction](#ui-elements) |
+| New entry: *metageneration* | [Word list](#word-list) |
+
+<a id="whats-new--20201130"></a>
+
+<a id="whats-new--november-30u002c-2020"></a>
+
+#### November 30, 2020
+
+| New guidance or change | Page |
+| --- | --- |
+| Updated content-navigation titles for some style guide pages. | Content navigation |
+| Clarified guidance for *just*. | [Word list](#word-list) |
+| New entry: *time to live* | [Word list](#word-list) |
+
+<a id="whats-new--20201123"></a>
+
+<a id="whats-new--november-23u002c-2020"></a>
+
+#### November 23, 2020
+
+| New guidance or change | Page |
+| --- | --- |
+| Changed guidance about fictitious email addresses. | [Example domains and names](#examples) |
+| New entry: *holiday, the holidays* | [Word list](#word-list) |
+
+<a id="whats-new--20201116"></a>
+
+<a id="whats-new--november-16u002c-2020"></a>
+
+#### November 16, 2020
+
+| New guidance or change | Page |
+| --- | --- |
+| In examples, use fictional street addresses. | [Example domains and names](#examples) |
+| Clarified guidance for *functionality*. | [Word list](#word-list) |
+
+<a id="whats-new--20201109"></a>
+
+<a id="whats-new--november-9u002c-2020"></a>
+
+#### November 9, 2020
+
+| New guidance or change | Page |
+| --- | --- |
+| Format folder and directory names in code font. | [Code in text](#code-in-text) |
+
+<a id="whats-new--20201102"></a>
+
+<a id="whats-new--november-2u002c-2020"></a>
+
+#### November 2, 2020
+
+| New guidance or change | Page |
+| --- | --- |
+| Avoid double negatives. | [Writing accessible documentation](#accessibility) |
+| Expanded guidance about how to write around unavoidable references to non-inclusive words. | [Writing inclusive documentation](#inclusive-documentation) |
+
+<a id="whats-new--20201019"></a>
+
+<a id="whats-new--october-19u002c-2020"></a>
+
+#### October 19, 2020
+
+We're changing the format of the updates to this page, to focus on what each
+change is rather than where it is.
+
+| New guidance or change | Page |
+| --- | --- |
+| You can use an asterisk to indicate emphasis in Markdown. Don't use a double underscore to indicate bold in Markdown. | [Text-formatting summary](#text-formatting) |
+
+<a id="whats-new--20201012"></a>
+
+<a id="whats-new--october-12u002c-2020"></a>
+
+#### October 12, 2020
+
+| New guidance or change | Page |
+| --- | --- |
+| Use *enable* or *turn on* consistently. | [Word list](#word-list--enable) |
+| All pages about linking are now grouped together under a new **Linking** heading. | Content navigation |
+| Don't link to the same document multiple times in close succession. | [Cross-references](https://developers.google.com/style/cross-references#references-to-other-documents) |
+| For subsections, use the phrase *in the following sections*. | [Headings and titles](https://developers.google.com/style/headings#referring-to-subsections) |
+| Don't create your own abbreviations. | [Abbreviations](#abbreviations--creating-abbreviations) |
+| Don't reveal personally identifiable information. | [Example domains and names](#examples) |
+| For exponents, use standard mathematical notation. | [Numbers](#numbers--exponents) |
+| Expanded guidance about expressing times of day. | [Dates and times](#dates-times--expressing-times) |
+| New entries: *aka*, *America*, *-aware*, *Black Friday*, *Cyber Monday*, *leverage*, *possible*, *postmortem* | [Word list](#word-list) |
+
+<a id="whats-new--20201005"></a>
+
+<a id="whats-new--october-5u002c-2020"></a>
+
+#### October 5, 2020
+
+- In the [word list](#word-list), updated guidance about
+  the term *disable*.
+
+<a id="whats-new--20200928"></a>
+
+<a id="whats-new--september-28u002c-2020"></a>
+
+#### September 28, 2020
+
+- On the [Capitalization](#capitalization) page, expanded
+  guidance about using consistent capitalization.
+- On the [Procedures](#procedures) page, added guidance
+  about including keyboard shortcuts in procedures.
+
+<a id="whats-new--20200921"></a>
+
+<a id="whats-new--september-21u002c-2020"></a>
+
+#### September 21, 2020
+
+- On the [Figures and other images](#images) page, added
+  guidance about using image maps.
+- On the [Text-formatting summary](#text-formatting)
+  page, added guidance about using a variable to indicate a range of version
+  numbers, and clarified guidance about using `<em>`.
+- Added entries to the [word list](#word-list):
+  *distributed denial-of-service (DDoS)*, *UTF*.
+
+<a id="whats-new--20200914"></a>
+
+<a id="whats-new--september-14u002c-2020"></a>
+
+#### September 14, 2020
+
+- On the [Figures and other images](#images) page and the
+  [Tables](#tables) page, added guidance about introductory
+  sentences.
+- In the [word list](#word-list), clarified the
+  distinction between terms to never use and terms to avoid when possible. Also
+  added icons for both kinds of terms.
+
+<a id="whats-new--20200907"></a>
+
+<a id="whats-new--september-7u002c-2020"></a>
+
+#### September 7, 2020
+
+- On the [Writing for a global audience](#translation)
+  page, clarified guidance about using simple verbs.
+- Added entries to the [word list](#word-list):
+  *single pane of glass*, *will*.
+
+<a id="whats-new--20200831"></a>
+
+<a id="whats-new--august-31u002c-2020"></a>
+
+#### August 31, 2020
+
+- On the [Figures and other images](#images)
+  page, clarified guidance about using SVG.
+
+<a id="whats-new--20200824"></a>
+
+<a id="whats-new--august-24u002c-2020"></a>
+
+#### August 24, 2020
+
+- Added an entry to the [word list](#word-list):
+  *see*.
+
+<a id="whats-new--20200810"></a>
+
+<a id="whats-new--august-10u002c-2020"></a>
+
+#### August 10, 2020
+
+- On the [UI elements and interaction](#ui-elements)
+  page, added guidance about when to use *in* and when to use *on*.
+- On the [Making headings into link
+  targets](#headings-targets) page, added guidance about changing anchors, and reorganized the
+  page.
+
+<a id="whats-new--20200803"></a>
+
+<a id="whats-new--august-3u002c-2020"></a>
+
+#### August 3, 2020
+
+- On the [Dates and times](#dates-times) page, added
+  guidance about avoiding referring to seasons of the year.
+- In the [word list](#word-list), added more suggested
+  alternatives for *master* and *slave*.
+
+<a id="whats-new--20200727"></a>
+
+<a id="whats-new--july-27u002c-2020"></a>
+
+#### July 27, 2020
+
+- Added entries to the [word list](#word-list):
+  *grandfathered*, *graylist*, *greylist*, *multi-regional*,
+  *Unix epoch time*. Expanded and clarified guidance for the
+  *blacklist*, *master*, and *multi-region* entries.
+
+<a id="whats-new--20200720"></a>
+
+<a id="whats-new--july-20u002c-2020"></a>
+
+#### July 20, 2020
+
+- Reorganized and updated the [Other editorial
+  resources](https://developers.google.com/style/resources) page, and moved that page to near the top of the left nav.
+- On the [Figures and other images](#images) page, added
+  links to Web Content Accessibility Guidelines (WCAG) resources.
+- On the [Procedures](#procedures) page, added guidance
+  about introducing a procedure with an imperative.
+- Added entries to the [word list](#word-list):
+  *bar*, *baz*, *canceled*, *foo*, *like*, *such
+  as*.
+- As of this week, we've stopped including small-detail changes in this
+  page. For example, when we corrected a typo in the style guide in the past, we
+  used to mention that correction on this page, but we're no longer mentioning
+  minor changes that have no effect on guidance.
+
+<a id="whats-new--20200713"></a>
+
+<a id="whats-new--july-13u002c-2020"></a>
+
+#### July 13, 2020
+
+- Renamed this summary-of-recent-changes page from "Release notes" to
+  "What's new," and moved it to near the top of the left nav.
+- In the [Pronouns](#pronouns) page, slightly updated the
+  guidance about using gender-neutral *they*.
+
+<a id="whats-new--20200629"></a>
+
+<a id="whats-new--june-29u002c-2020"></a>
+
+#### June 29, 2020
+
+- On several pages, reorganized guidance about formatting code and
+  command lines. In particular:
+  - Moved the guidance that was in a page about command-line terminology
+    to the [Documenting command-line syntax](#code-syntax) page,
+    and deleted the separate command-line terminology page.
+  - Moved the guidance about placeholders to a new [Formatting placeholders](#placeholders) page.
+  - Added cross-references among various code-related pages.
+- In the [word list](#word-list), corrected a typo.
+
+<a id="whats-new--20200622"></a>
+
+<a id="whats-new--june-22u002c-2020"></a>
+
+#### June 22, 2020
+
+- In the [word list](#word-list), changed guidance about
+  using *earlier* and *later* for a range of version numbers.
+
+<a id="whats-new--20200615"></a>
+
+<a id="whats-new--june-15u002c-2020"></a>
+
+#### June 15, 2020
+
+- On the [Capitalization](#capitalization) page and other
+  pages, made guidance consistent about capitalizing references to document
+  titles.
+- On the [Cross-references](#cross-references) page and
+  other pages, added guidance about not forcing a link to open in a new tab.
+- On the [Code in text](#code-in-text) page, added
+  guidance about when to put UI elements in code font.
+- Added an entry to the [word list](#word-list): *best
+  effort*.
+
+<a id="whats-new--20200608"></a>
+
+<a id="whats-new--june-8u002c-2020"></a>
+
+#### June 8, 2020
+
+- On the [Writing inclusive
+  documentation](#inclusive-documentation) page, added guidance about avoiding figurative language that
+  relates to the slaughter of animals.
+
+<a id="whats-new--20200601"></a>
+
+<a id="whats-new--june-1u002c-2020"></a>
+
+#### June 1, 2020
+
+- On the [Figures and other images](#images) page,
+  updated guidance about alt text, figure descriptions, and figure captions.
+- Reorganized the [HTML and semantic
+  tagging](#semantic-tagging) page, and added guidance about not using the [`br`
+  element](https://html.spec.whatwg.org/multipage/text-level-semantics.html#the-br-element) to change visual presentation. Also replaced a broken link to a
+  third-party site.
+- On the [Numbers](#numbers) page, added guidance about
+  avoiding using Roman numerals.
+- Added entries to the [word list](#word-list):
+  *currently*, *dash*, *NoOps*.
+
+<a id="whats-new--20200525"></a>
+
+<a id="whats-new--may-25u002c-2020"></a>
+
+#### May 25, 2020
+
+- On the [Numbers](#numbers) page, corrected two examples
+  to match our guidance.
+- On the [Abbreviations](#abbreviations) page, added
+  guidance about use of the term *acronym*.
+- Added entries to the [word list](#word-list):
+  *happiness*, *scale*, *service level agreement*, *service level
+  indicator*, *service level objective*. Also clarified guidance for
+  *app*.
+
+<a id="whats-new--20200518"></a>
+
+<a id="whats-new--may-18u002c-2020"></a>
+
+#### May 18, 2020
+
+- On the [Filenames and file types](#filenames) page,
+  added guidance about how to refer to a filename.
+- On the [Headings and titles](#headings) page, updated
+  the guidance about using code font in headings.
+- On the [UI elements and interaction](#ui-elements) page
+  and elsewhere, replaced the term *arrow notation* with *angle
+  bracket*.
+- On the [Procedures](#procedures) page, clarified the
+  guidance about stating the location of an action before stating the action.
+- Added entries to the [word list](#word-list):
+  *config*, *extract*, *unarchive*, *uncompress*,
+  *untar*. Also changed the guidance about *ssh* and related terms.
+  Also corrected a typo.
+
+<a id="whats-new--20200511"></a>
+
+<a id="whats-new--may-11u002c-2020"></a>
+
+#### May 11, 2020
+
+- Major change: On the [Code in text](#code-in-text)
+  page, changed guidance for placeholder styling. The new preferred style is
+  all-uppercase with underscore delimiters.
+- Also on the [Code in text](#code-in-text) page,
+  corrected guidance about how to mark up placeholders with Markdown.
+- On the [Figures and other images](#images) page,
+  corrected a typo.
+
+<a id="whats-new--20200504"></a>
+
+<a id="whats-new--may-4u002c-2020"></a>
+
+#### May 4, 2020
+
+- On the [Second person and first person](#person) page,
+  clarified guidance about when to use implicit *you*.
+- On the [UI elements and interaction](#ui-elements)
+  page, added guidance about *toggle*.
+- Added entries to the [word list](#word-list):
+  *dashboard*, *master*. Unrelatedly, fixed a longstanding word list
+  typo.
+
+<a id="whats-new--20200427"></a>
+
+<a id="whats-new--april-27u002c-2020"></a>
+
+#### April 27, 2020
+
+- On the [Documenting command-line syntax](#code-syntax)
+  page, added guidance about showing the output of commands.
+- On the [Example domains and names](#examples) page,
+  added example IPv6 addresses.
+- On the [Words as words](https://developers.google.com/style/words-as-words) page,
+  clarified guidance.
+- In the [word list](#word-list), clarified guidance for
+  the word *following*. Also clarified guidance for the various
+  *console* terms.
+- Added an entry to the [word list](#word-list):
+  *native*.
+
+<a id="whats-new--20200420"></a>
+
+<a id="whats-new--april-20u002c-2020"></a>
+
+#### April 20, 2020
+
+- On the [Units of measurement](#units-of-measure) page,
+  added guidance about using *per* when referring to rates.
+- On the [UI elements and interaction](#ui-elements)
+  page, clarified the distinction between text input and keyboard shortcuts, and
+  added guidance about when to use the `code` and
+  `kbd` elements. On the [Code in
+  text](#code-in-text) page, added text input to the list of things to put in code font.
+- On the [Accessible content](#accessibility) page, added
+  a recommendation to use a screen reader to test your documentation. Also added
+  guidance about introducing tables.
+- On the [Example domains and names](#examples) page,
+  changed the recommended example company names to avoid camel case.
+- On the [Code in text](#code-in-text) page and the [Colons](#colons) page, clarified and expanded guidance about how
+  to format a colon after text that's bold or in code font.
+- Added entries to the [word list](#word-list):
+  *key*, *key ring*, *per*, *vice versa*.
+
+<a id="whats-new--20200413"></a>
+
+<a id="whats-new--april-13u002c-2020"></a>
+
+#### April 13, 2020
+
+- On the [Lists](#lists) page, expanded the guidance
+  about when not to use sentence-ending punctuation.
+- Added cross-references between the [Abbreviations](#abbreviations) page and the [Headings and titles](#headings) page.
+- On the [Numbers](#numbers) page, clarified the
+  guidance about hyphens and en dashes in ranges of numbers.
+- On the [Units of measure](#units-of-measure) page,
+  clarified guidance about what counts as a unit.
+- In the [word list](#word-list), clarified guidance
+  about using *blacklist* and *whitelist*.
+- Added entries to the [word list](#word-list):
+  *ad tech*, *easily*, *fintech*, *NLU*, *quick*,
+  *quickly*.
+
+<a id="whats-new--20200406"></a>
+
+<a id="whats-new--april-6u002c-2020"></a>
+
+#### April 6, 2020
+
+- On the [Abbreviations](#abbreviations) page, clarified
+  guidance about using abbreviations in headings and titles.
+- On the [Text-formatting summary](#text-formatting)
+  page, added guidance on formatting mathematical variables.
+- On the [Link text](https://developers.google.com/style/link-text) page, added guidance
+  about abbreviations in link text.
+- On the [Dates and times](#dates-times) page, added
+  guidance about how to express time zones.
+- In the [word list](#word-list), clarified guidance
+  about *key-value pair*.
+- Added an entry to the [word list](#word-list):
+  *personally identifiable information (PII)*.
+
+<a id="whats-new--20200330"></a>
+
+<a id="whats-new--march-30u002c-2020"></a>
+
+#### March 30, 2020
+
+- On the [Hyphens](#hyphens) page, clarified guidance
+  about ranges of numbers.
+- In the [UI elements and interaction](#ui-elements)
+  page, changed guidance for formatting input for text boxes to recommend using
+  the HTML `code` element instead of the
+  `kbd` element.
+
+<a id="whats-new--20200323"></a>
+
+<a id="whats-new--march-23u002c-2020"></a>
+
+#### March 23, 2020
+
+- On the [Capitalization](#capitalization) page, made the
+  following changes:
+  - Removed outdated guidance about capitalization after a run-in heading.
+    For the current guidance, see "Description lists that use run-in headings" on
+    the [Lists](#lists) page.
+  - Clarified guidance about avoiding all-uppercase and camel case.
+  - Clarified guidance about avoiding lowercase at the start of a
+    sentence.
+  - Clarified guidance about capitalization when referring to documents by
+    title in various contexts.
+- Expanded and reorganized the [Code in
+  text](#code-in-text) page. In particular:
+  - Expanded the guidance about documenting placeholders.
+  - Expanded the list of items to put in code font.
+- On the [Commas](#commas) page, added guidance about
+  using a comma with *because*.
+- On the [Dashes](#dashes) page, added guidance about not
+  using en dashes.
+- Added entries to the [word list](#word-list):
+  *datastore*, *tutorial*. Also added guidance (in the entry for
+  *documentation*) about phrases like *in this document*.
+
+<a id="whats-new--20200302"></a>
+
+<a id="whats-new--march-2u002c-2020"></a>
+
+#### March 2, 2020
+
+- On the [Capitalization](#capitalization) page, added
+  guidance about using sentence case for labels and callouts.
+- On the [Cross-references](#cross-references) page,
+  added guidance about how to format the titles of web series.
+- Added an entry to the [word list](#word-list): *web
+  server*.
+
+<a id="whats-new--20200224"></a>
+
+<a id="whats-new--february-24u002c-2020"></a>
+
+#### February 24, 2020
+
+- On the [Resources](https://developers.google.com/style/resources) page, added links to the
+  Kubernetes glossary and the Kubernetes style guide.
+- On the [Lists](#lists) page, added guidance about how
+  and when to use a bulleted list with run-in headings.
+- Updated the [word list](#word-list) to refer to
+  *Google Cloud* instead of *GCP*, in accord with a recent name change.
+  In particular, changed entries for *GCP*, *GCP Console*, and *GCP
+  project ID* to *Google Cloud*, *Google Cloud Console*, and
+  *Google Cloud project ID*, respectively.
+- Added an entry to the [word list](#word-list): *GKE
+  node*. Also updated guidance for *GKE*.
+- In the left nav, moved the word list and the product-names page to near
+  the top.
+- On the [Plurals in parentheses](https://developers.google.com/style/plurals-parentheses)
+  page, added guidance about using the phrase *one or more*.
+
+<a id="whats-new--20200113"></a>
+
+<a id="whats-new--january-13u002c-2020"></a>
+
+#### January 13, 2020
+
+- In the [Writing for a global audience](#translation)
+  page, fixed a broken link.
+- In the [Code in text](#code-in-text) page, added
+  *environment variable names* to the list of items to put in code font.
+- Added entries to the [word list](#word-list):
+  *A/B testing*, *BGP*.
+
+<a id="whats-new--20200106"></a>
+
+<a id="whats-new--january-6u002c-2020"></a>
+
+#### January 6, 2020
+
+- Added entries to the [word list](#word-list):
+  *DevOps*, *ECMP*, *GFE*, *NoSQL*, *PostgreSQL*. Also
+  updated guidance on *virtual machine (VM) instance*.
+
+<a id="whats-new--20191223"></a>
+
+<a id="whats-new--december-23u002c-2019"></a>
+
+#### December 23, 2019
+
+- Expanded and clarified the [Writing for a
+  global audience](#translation) page.
+- Renamed the "Typographical conventions" page to "[Text-formatting summary](#text-formatting)."
+- On the [Tables](#tables) page, clarified guidance
+  about punctuation in table column heads.
+- On the [Filenames and file types](#filenames) page,
+  added *.jar* to the list of filename extensions.
+- On the [UI elements](#ui-elements) page and the word
+  list, clarified guidance about when to use *box* and *field* to refer
+  to text boxes.
+
+<a id="whats-new--20191216"></a>
+
+<a id="whats-new--december-16u002c-2019"></a>
+
+#### December 16, 2019
+
+- On the [Abbreviations](#abbreviations) page, expanded
+  guidance about when to spell out abbreviations.
+- On the [conventions page](https://developers.google.com/style/typographical-conventions),
+  fixed two broken links.
+- Added entries to the [word list](#word-list):
+  *as*, *big-endian*, *google.dev*, *little-endian*.
+
+<a id="whats-new--20191209"></a>
+
+<a id="whats-new--december-9u002c-2019"></a>
+
+#### December 9, 2019
+
+- On the [Tables](#tables) page, added guidance about
+  sorting tables, and clarified guidance about when to use tables.
+- Reverted last week's changes about using *earlier* and
+  *later* instead of *lower* and *higher*; those changes need
+  further refining.
+
+<a id="whats-new--20191202"></a>
+
+<a id="whats-new--december-2u002c-2019"></a>
+
+#### December 2, 2019
+
+- On the [Cross-references](#cross-references)
+  page, added guidance about using the word *about* in cross-references,
+  rather than *on*.
+- Added entries to the [word list](#word-list): *about
+  versus on*, *for instance*, *SLA*. Also changed guidance about
+  using *earlier* and *later* for a range of version numbers.
+
+<a id="whats-new--20191125"></a>
+
+<a id="whats-new--november-25u002c-2019"></a>
+
+#### November 25, 2019
+
+- On the [UI elements and interaction](#ui-elements)
+  page, added guidance about focusing on the task rather than on UI elements as
+  such.
+- On various pages, added cross-references linking to other pages.
+
+<a id="whats-new--20191118"></a>
+
+<a id="whats-new--november-18u002c-2019"></a>
+
+#### November 18, 2019
+
+- Added a new [Typographic
+  conventions](https://developers.google.com/style/typographical-conventions) page.
+- Reorganized the [Units of
+  measurement](#units-of-measure) page. Also added guidance about Kelvin temperatures.
+- On the [Cross-references](#cross-references) page,
+  clarified guidance about cross-references that aren't links.
+- On the [UI elements and interaction](#ui-elements)
+  page, added information about expander arrows.
+- Added a term to the [word list](#word-list):
+  *slice and dice*.
+
+<a id="whats-new--20191111"></a>
+
+<a id="whats-new--november-11u002c-2019"></a>
+
+#### November 11, 2019
+
+- On the [Figures and other images](#images) page,
+  clarified guidance about figure descriptions.
+- On the [Units of measurement](#units-of-measure) page,
+  added guidance about ranges of numbers.
+- On the [Product names](#product-names) page, clarified
+  guidance about using full trademarked product names.
+- Added terms to the [word list](#word-list):
+  *billing charges*, *port*, *repo*, *UI*.
+
+<a id="whats-new--20191028"></a>
+
+<a id="whats-new--october-28u002c-2019"></a>
+
+#### October 28, 2019
+
+- Added guidance about what to do if there's [more than one way to complete a
+  task](#procedures--same-task).
+- Added terms to the [word list](#word-list):
+  *just*, *multi-tenancy*.
+
+<a id="whats-new--20191021"></a>
+
+<a id="whats-new--october-21u002c-2019"></a>
+
+#### October 21, 2019
+
+- Added a page about [formatting
+  words as words](https://developers.google.com/style/formatting-words-as-words).
+- Added a page about [formatting key
+  terms](https://developers.google.com/style/formatting-key-terms).
+- Changed the guidance about how to capitalize a [UI element](#ui-elements) name when the label is
+  all-uppercase in the UI.
+- Added guidance about [explaining
+  placeholders](https://developers.google.com/style/code-in-text#explain-placeholders).
+- Updated the list of [items to put in
+  code font](#code-in-text--code).
+- Clarified the guidance about using official [capitalization](#capitalization) for names of products.
+- Changed the title of the [style guide landing
+  page](#style).
+- Deleted the page about URLs in links, and moved the material that was on
+  that page to the [cross-references](#cross-references)
+  page. Also added some examples to that page.
+- Removed unnecessary explanations of the icons that indicate that a
+  guideline applies only in Android documentation or only in Google Cloud Platform
+  documentation.
+- Added examples to the [footnotes](#footnotes)
+  page.
+- Improved a link on the [Accessible
+  content](#accessibility) page.
+- Added terms to the [word list](#word-list):
+  *BMaaS*, *US*.
+
+<a id="whats-new--20191008"></a>
+
+<a id="whats-new--october-8u002c-2019"></a>
+
+#### October 8, 2019
+
+- Reorganized the page about [notices](#notices), and
+  added examples.
+- Cleaned up some minor formatting and phrasing issues in the [word list](#word-list).
+- Corrected a typo on the [Accessible
+  content](#accessibility) page.
+- Changed the title of the style guide to use sentence case.
+
+<a id="whats-new--20190930"></a>
+
+<a id="whats-new--september-30u002c-2019"></a>
+
+#### September 30, 2019
+
+- Added a term to the [word list](#word-list):
+  *PoP*.
+
+<a id="whats-new--20190923"></a>
+
+<a id="whats-new--september-23u002c-2019"></a>
+
+#### September 23, 2019
+
+- Added a page about [footnotes](#footnotes).
+- Reorganized the [Other editorial resources](https://developers.google.com/style/resources)
+  page.
+
+<a id="whats-new--20190916"></a>
+
+<a id="whats-new--september-16u002c-2019"></a>
+
+#### September 16, 2019
+
+- Added a page about [using material from
+  other sources](#other-sources).
+- Added guidance about [using
+  *a* and *an* with product names](#product-names--shortening).
+- Removed contradictory guidance about quotation marks with [link text](https://developers.google.com/style/link-text).
+- Clarified guidance about [periods
+  with abbreviations](#abbreviations--periods).
+- Clarified guidance about [capitalization
+  following a colon](#capitalization).
+- Moved the list of UI-related verbs from the UI elements page to the [word list](#word-list).
+- Also added other terms to the [word list](#word-list):
+  *cons*, *Gbps*, *Kbps*, *Mbps*, *pros*. Also updated
+  the entry for *service*.
+
+<a id="whats-new--20190624"></a>
+
+<a id="whats-new--june-24u002c-2019"></a>
+
+#### June 24, 2019
+
+- Added guidance about US currency to the [Numbers](#numbers) page.
+
+<a id="whats-new--20190617"></a>
+
+<a id="whats-new--june-17u002c-2019"></a>
+
+#### June 17, 2019
+
+- Added further guidance about accessibility to the [Accessible content](#accessibility) page and other related
+  pages.
+- Added guidance about graphically violent terms to the [Writing inclusive documentation](#inclusive-documentation)
+  page.
+- Added guidance about the term *wheelchair-bound* to the [Writing inclusive documentation](#inclusive-documentation)
+  page.
+- Added guidance about currency to the [Units of measure](#units-of-measure) page.
+- Added an example showing how to [use
+  Markdown to mark up images](#images).
+- Added terms to the [word list](#word-list):
+  *disclosure triangle*, *hamburger menu*, *kebab menu*,
+  *outpost*, *STONITH*. Also expanded guidance about *should*.
+
+<a id="whats-new--20190520"></a>
+
+<a id="whats-new--may-20u002c-2019"></a>
+
+#### May 20, 2019
+
+- Added a new page about [avoiding excessive claims](#excessive-claims).
+- Added material about file types to the [Filenames and file types](#filenames) page.
+- Added terms to the [word list](#word-list): *CSV*,
+  *traditional*, *YAML*.
+
+<a id="whats-new--20190429"></a>
+
+<a id="whats-new--april-29u002c-2019"></a>
+
+#### April 29, 2019
+
+- Replaced the guidelines on the [Accessible content](#accessibility) page with a new set of more
+  relevant guidelines.
+- Added a term to the [word list](#word-list):
+  *execute*.
+- Added material about Linux signals on the [Command-line terminology](https://developers.google.com/style/command-line-terminology) page, and
+  updated related entries in the word list.
+
+<a id="whats-new--20190422"></a>
+
+<a id="whats-new--april-22u002c-2019"></a>
+
+#### April 22, 2019
+
+- Added a term to the [word list](#word-list):
+  *method*. Also added to the guidance for *Cloud*.
+
+<a id="whats-new--20190415"></a>
+
+<a id="whats-new--april-15u002c-2019"></a>
+
+#### April 15, 2019
+
+- Added more guidance about digit-group separators on the [Numbers](#numbers) page.
+- Fixed a broken link on the [Numbers](#numbers)
+  page, and another on the [Writing for a global
+  audience](#translation) page.
+- Updated guidance about example names on the [Example domains and names](#examples) page.
+- Added guidance about removing personally identifying information from
+  screenshots on the [Figures and other images](#images)
+  page.
+- Added guidance about using code font for names of command-line utilities
+  on the [Code in text](#code-in-text) page.
+- Added guidance about short and long versions of words like *app* and
+  *demo* on the [Abbreviations](#abbreviations) page.
+
+<a id="whats-new--20190304"></a>
+
+<a id="whats-new--march-4u002c-2019"></a>
+
+#### March 4, 2019
+
+- Clarified guidance about person-first and identity-first language on the
+  [Writing inclusive documentation](#inclusive-documentation)
+  page.
+- Added guidance about passive voice on the [Active voice](#voice) page.
+- Added more guidance about links on the [Linking to other sites](https://developers.google.com/style/links-external) page.
+- Added terms to the [word list](#word-list):
+  *guru*, *spin up*. Also revised the guidance for *Container
+  Engine*.
+
+<a id="whats-new--20190225"></a>
+
+<a id="whats-new--february-25u002c-2019"></a>
+
+#### February 25, 2019
+
+- Added extensive new material to the [inclusive documentation](#inclusive-documentation) page.
+- Updated the [word list](#word-list):
+  - Clarified guidance for *argument* and *flag*.
+  - Added terms: *allows you to*, *desire*, *drop-down*,
+    *wish*.
+  - Removed the recommendations against using *enable* and
+    *disable*, and clarified guidance for those terms.
+
+<a id="whats-new--20190204"></a>
+
+<a id="whats-new--february-4u002c-2019"></a>
+
+#### February 4, 2019
+
+- Added terms to the [word list](#word-list):
+  *add-in*, *add-on*.
+- Added guidelines about avoiding [repetitive procedures](#procedures--repetitive-procedures).
+
+<a id="whats-new--20190128"></a>
+
+<a id="whats-new--january-28u002c-2019"></a>
+
+#### January 28, 2019
+
+- Updated the [word list](#word-list):
+  - Changed guidance for *above* and *below*.
+  - Added terms: *gcloud*, *ML*.
+  - Updated guidance for *&*.
+  - Improved alphabetization.
+- Fixed a broken link in the [Lists](#lists) page.
+
+<a id="whats-new--20190114"></a>
+
+<a id="whats-new--january-14u002c-2019"></a>
+
+#### January 14, 2019
+
+- Added terms to the [word list](#word-list):
+  *comprise*, *console*, *fail over*, *GCP Console*, *Google
+  Cloud Platform Console*, *populate*, *surface*.
+- Reorganized the [UI elements](#ui-elements) page,
+  and added guidance and examples.
+- Added Markdown information to the [URLs for
+  images](https://developers.google.com/style/img-elements) page.
+- Clarified guidance about end punctuation for lists on the [Periods](#periods) page.
+- Clarified guidance about nonbreaking spaces in the [Units of measurement](#units-of-measure) page.
+
+<a id="whats-new--20181203"></a>
+
+<a id="whats-new--december-3u002c-2018"></a>
+
+#### December 3, 2018
+
+- Added a term to the [word list](#word-list):
+  *lifetime*.
+- Elaborated on our guidance in the [Closed compounds and prefixes](https://developers.google.com/style/word-list#compounds) section of
+  the word list.
+- Added a new page about [command-line terminology](https://developers.google.com/style/command-line-terminology) for gcloud
+  and Linux.
+- Added a section on [straight and
+  curly quotation marks](#quotation-marks--straight-and-curly-quotation-marks) to the Quotation marks page.
+- Added guidance about avoiding overuse of bolding to the [Procedures](#procedures) page.
+- Added guidance about the phrases *hold the pointer over* and
+  *point to* to the [UI elements and
+  interaction](#ui-elements) page.
+
+<a id="whats-new--20181029"></a>
+
+<a id="whats-new--october-29u002c-2018"></a>
+
+#### October 29, 2018
+
+- Added terms to the [word list](#word-list):
+  *blacklist*, *whitelist*.
+- Reorganized the [dates and times](#dates-times) page,
+  and added guidance about times and about ISO 8601.
+
+<a id="whats-new--20181008"></a>
+
+<a id="whats-new--october-8u002c-2018"></a>
+
+#### October 8, 2018
+
+- Added recommendations for [example company names](#examples--example-company-names).
+- Added guidance about using monospace formatting to the [UI elements and interaction](#ui-elements) page.
+- Added guidance about giving the location of an action in a procedure
+  to the [Procedures](#procedures) page.
+
+<a id="whats-new--20180918"></a>
+
+<a id="whats-new--september-18u002c-2018"></a>
+
+#### September 18, 2018
+
+- Added terms to the [word list](#word-list): *GKE*,
+  *POJO*.
+
+<a id="whats-new--20180904"></a>
+
+<a id="whats-new--september-4u002c-2018"></a>
+
+#### September 4, 2018
+
+- Added guidance about cases where [filenames aren't under your
+  control](#filenames--other-exceptions).
+- Added a term to the [word list](#word-list):
+  *big data*.
+
+<a id="whats-new--20180813"></a>
+
+<a id="whats-new--august-13u002c-2018"></a>
+
+#### August 13, 2018
+
+- Added guidance about [avoiding internet slang](#abbreviations--dont-use).
+- Added terms to the [word list](#word-list):
+  *multi-cloud*, *RTFM*, *tl;dr*, *ymmv*.
+
+<a id="whats-new--20180730"></a>
+
+<a id="whats-new--july-30u002c-2018"></a>
+
+#### July 30, 2018
+
+- Added terms to the [word list](#word-list):
+  *disaster recovery*, *file system*, *high availability*,
+  *high performance computing*, *OS*, *tarball*, *tar
+  file*.
+- Updated guidance in the word list for *on-premises*.
+- Rephrased a paragraph of the [landing page](#style).
+
+<a id="whats-new--20180625"></a>
+
+<a id="whats-new--june-25u002c-2018"></a>
+
+#### June 25, 2018
+
+- Added terms to the [word list](#word-list):
+  *IP*, *SAP*.
+
+<a id="whats-new--20180619"></a>
+
+<a id="whats-new--june-19u002c-2018"></a>
+
+#### June 19, 2018
+
+- Added a term to the [word list](#word-list):
+  *trojan*.
+- Improved the organization, formatting, and phrasing of the [filenames page](https://developers.google.com/style/file-names).
+
+<a id="whats-new--20180612"></a>
+
+<a id="whats-new--june-12u002c-2018"></a>
+
+#### June 12, 2018
+
+- Added terms to the [word list](#word-list):
+  *appendix*, *index*, *matrix*.
+
+<a id="whats-new--20180521"></a>
+
+<a id="whats-new--may-21u002c-2018"></a>
+
+#### May 21, 2018
+
+- Changed the guidance about [how to link to Android
+  reference documentation](#code-in-text--linking-api-terms-in-android).
+- Added information about [using *the* with
+  product names](https://developers.google.com/style/product-names#using-the-with-product-names).
+- Added a note about indefinite and casual numbers to the [Numbers](#numbers) page.
+- Fixed a broken link on the [authorial tone](#tone)
+  page.
+
+<a id="whats-new--20180514"></a>
+
+<a id="whats-new--may-14u002c-2018"></a>
+
+#### May 14, 2018
+
+- Added terms to the [word list](#word-list):
+  *multi-cluster*, *via*. Clarified guidance on *firewalls*.
+
+<a id="whats-new--20180402"></a>
+
+<a id="whats-new--april-2u002c-2018"></a>
+
+#### April 2, 2018
+
+- Added a page about [writing inclusive documentation](#inclusive-documentation).
+- Added terms to the [word list](#word-list):
+  *at scale*, *IAM*, *RFC*.
+
+<a id="whats-new--20180326"></a>
+
+<a id="whats-new--march-26u002c-2018"></a>
+
+#### March 26, 2018
+
+- Added information about [`alt` text for
+  decorative images](#images--text-associated-with-images) to the images page.
+- Added terms to the [word list](#word-list):
+  *fill in*, *fill out*, *microservices*.
+
+<a id="whats-new--20180226"></a>
+
+<a id="whats-new--february-26u002c-2018"></a>
+
+#### February 26, 2018
+
+- Added a term to the [word list](#word-list):
+  *ad hoc*.
+
+<a id="whats-new--20180212"></a>
+
+<a id="whats-new--february-12u002c-2018"></a>
+
+#### February 12, 2018
+
+- Added terms to the [word list](#word-list):
+  *Cloud*, *Container Engine*, *ephemeral external IP address*,
+  *firewalls*, *GCP*, *GCP project ID*, *internal IP
+  address*, *NAT*, *network IP address*, *non-key*,
+  *on-premises*, *persistent disk*, *project*, *RDP*,
+  *static external IP address*, *subnet*, *table name*,
+  *textbox*, *turn on*.
+- Added information about closed compounds and prefixes to the [word list](https://developers.google.com/style/word-list#compounds).
+
+<a id="whats-new--20180129"></a>
+
+<a id="whats-new--january-29u002c-2018"></a>
+
+#### January 29, 2018
+
+- Added terms to the [word list](#word-list):
+  *alpha*, *autohealing*, *autoscaling*, *autotagging*,
+  *bare metal*, *beta*, *CLI*, *Cloud SDK*, *colocate*,
+  *CPU*, *curated roles*, *data center campus*, *data
+  cleaning*, *data source*, *ETL*, *IaaS*, *impact*,
+  *ingest*, *k8s*, *key pair*, *PaaS*, *preemptible*,
+  *regex*, *SaaS*, *screenshot*, *sign-in*, *sign into*,
+  *sign-out*, *sub-command*, *v*, *virtual machine
+  instance*.
+- Added a section on placeholder variables to the [code in text](https://developers.google.com/style/code-in-text#placeholder-variables) page.
+
+<a id="whats-new--20180108"></a>
+
+<a id="whats-new--january-8u002c-2018"></a>
+
+#### January 8, 2018
+
+- Added terms to the [word list](#word-list):
+  *argument*, *backup*, *dataflow*, *flag*, *higher*,
+  *later*, *long press*, *multi-region*, *option*,
+  *slave*, *touch & hold*, *transpile*, *wildcard*.
+- Added an item (DNS record types) to the list of [items to put
+  in code font](#code-in-text--some-specific-items-to-put-in-code-font).
+
+<a id="whats-new--20171211"></a>
+
+<a id="whats-new--december-11u002c-2017"></a>
+
+#### December 11, 2017
+
+- Added material on [numerical
+  dimensions](#numbers--dimensions).
+- Added material about [possessives of
+  company and product names](https://developers.google.com/style/possessives#company--and-product-name-possessives).
+- Added terms to the [word list](#word-list):
+  *docset*, *documentation*, *startup*, *sync*.
+
+<a id="whats-new--20171113"></a>
+
+<a id="whats-new--november-13u002c-2017"></a>
+
+#### November 13, 2017
+
+- Changed guidelines on [end punctuation for
+  list items](#lists--capitalization-and-end-punctuation).
+- Added alphabetical headings to the [word
+  list](#word-list), so you can now jump to a specific letter using the in-page TOC.
+- Updated [word list](#word-list) entries for
+  *error-prone*, *real time*, and *Search*.
+- Added terms to the [word list](#word-list):
+  *dataset*, *legacy*, *nonfatal*.
+
+<a id="whats-new--20171023"></a>
+
+<a id="whats-new--october-23u002c-2017"></a>
+
+#### October 23, 2017
+
+- Added a new page about [plurals in
+  parentheses](https://developers.google.com/style/plurals-parentheses).
+- Added a link to the C++ coding style guide, in the [Code style guides](#code-samples--coding) section.
+
+<a id="whats-new--20170904"></a>
+
+<a id="whats-new--september-4u002c-2017"></a>
+
+#### September 4, 2017
+
+- Added several more terms to the [word list](#word-list):
+  *check*, *checkbox*, *clear*, *deselect*, *open
+  source*, *select*, *target*, *uncheck*, *Unix-like*,
+  *unselect*.
+- Added material to the [Trademarks](#trademarks) page
+  about using trademarks as adjectives. Also added links to a couple of other
+  trademark-related pages from Google.
+- Removed Android-specific material from the [Headings](#headings) page, to bring Android doc guidelines more
+  into line with non-Android guidelines. Also, added notes about things not to
+  include in headings.
+- Removed Android label from material on the [Tables](#tables) page, to make that material also apply to
+  non-Android docs.
+- Added mention of degrees and percents to the [Units of measurement](#units-of-measure) page.
+
+<a id="whats-new--20170828"></a>
+
+<a id="whats-new--august-28u002c-2017"></a>
+
+#### August 28, 2017
+
+- Added 160+ terms to the [word list](#word-list). (Too
+  many to list here.)
+- Added recommendations for image size and placement to the [Images](#images) page.
+- Moved material about Javadoc links to the [Code in text](#code-in-text) page.
+
+<a id="whats-new--20170821"></a>
+
+<a id="whats-new--august-21u002c-2017"></a>
+
+#### August 21, 2017
+
+- Added new [File names](https://developers.google.com/style/file-names) page.
+- Added new [Units of measurement](#units-of-measure)
+  page.
+- Changed "[internet](#word-list--internet)" to
+  lowercase.
+- Added material about line lengths and Javadoc-style links to the [HTML formatting](#html-formatting) page.
+
+<a id="whats-new--20170731"></a>
+
+<a id="whats-new--july-31u002c-2017"></a>
+
+#### July 31, 2017
+
+- Added new [Headings and titles](#headings) page, mostly
+  focused on Android documentation.
+- Added information about [tables in Android docs](https://developers.google.com/style/tables#a-few-more-notes-about-tables) to
+  the Tables page.
+
+<a id="whats-new--20170710"></a>
+
+<a id="whats-new--july-10u002c-2017"></a>
+
+#### July 10, 2017
+
+- Added words to the [word list](#word-list): *deep
+  linking*, *jank*, *wake lock*.
+
+<a id="whats-new--20170608"></a>
+
+<a id="whats-new--june-8u002c-2017"></a>
+
+#### June 8, 2017
+
+- Public release.
+
+---
+
+<a id="philosophy"></a>
+
+### Philosophy of this guide
+
+*Source: <https://developers.google.com/style/philosophy>*
+
+This document discusses some of the principles and philosophy behind this
+style guide.
+
+<a id="philosophy--purpose"></a>
+
+<a id="philosophy--intended-purpose"></a>
+
+#### Intended purpose
+
+This style guide codifies and records our style decisions and describes our
+house style. The guide doesn't claim to be objectively correct.
+
+This guide is *not* intended to do the following:
+
+- Provide an industry documentation standard.
+- Compete with other well-known style guides.
+- Replace another style guide that you already follow.
+- Provide a complete set of basic writing guidelines.
+- Provide legal advice. For legal advice, consult a lawyer.
+
+> [!NOTE]
+> **Note**: Two disclaimers:
+> - The guidance in this style guide doesn't limit the changes that Google can
+>   make to its documentation.
+> - If you don't read a given guideline, then you are still responsible for
+>   behaving ethically and lawfully with regard to documentation.
+
+<a id="philosophy--reasons-for-guidelines"></a>
+
+<a id="philosophy--explanation-of-reasons-for-guidelines"></a>
+
+#### Explanation of reasons for guidelines
+
+We generally don't explain the reasoning behind most of our guidelines. We
+have a couple of reasons for that:
+
+- Many of our decisions are driven by accessibility, localization,
+  globalization, and ease of understanding. Giving those reasons as explanations
+  everywhere they apply would be repetitive.
+- Often, a given guideline is one good option among several; in those cases,
+  we sometimes just chose one option for consistency.
+- Too much explanation can clutter up a page. Readers most often want a
+  brief answer to a specific question, rather than a detailed explanation.
+
+That said, we recognize that it's sometimes useful to know why we made a
+given choice, so we've started to include occasional explanations in the [What's new](#whats-new) page.
+
+---
+
+## Key resources
+
+<a id="word-list"></a>
+
+### Word list
+
+*Source: <https://developers.google.com/style/word-list>*
+
+> [!NOTE]
+> **Note**: This document includes references to potentially disrespectful
+> or offensive terms. These terms are listed here to provide usage
+> guidance and alternative terms.
+
+<a id="word-list--overview"></a>
+
+This word list covers style and usage guidelines that are specific to developer documentation.
+
+If the term that you're looking for isn't on this list, check our other
+[editorial resources](#style--editorial-resources), including our preferred
+dictionary,
+[Merriam-Webster](https://www.merriam-webster.com/). If there are multiple spellings in
+the Merriam-Webster word entry, use the first form listed, which is the most common spelling. For
+example, in the
+[entry for *cancel*](https://www.merriam-webster.com/dictionary/canceled),
+the first form listed for the past tense is *canceled*, indicating that it's more common than
+*cancelled*.
+
+If you're looking for a technical definition, then it's often a good idea to check the
+authoritative documentation on the topic.
+
+Terminology decisions, including how and when to define or contextualize
+terms, often require judgments based on factors like your product area,
+your audience, and prevailing convention. Here are some other pages of this
+guide that can help you make those types of judgments:
+
+- [Jargon](#jargon)
+- [Inclusive language](#inclusive-documentation)
+- [Write for a global audience](#translation)
+- [Hyphens](#hyphens)
+- [Capitalization](#capitalization)
+
+As always, it's fine to deviate from our guidance if that serves your readers
+better. For more information, see [Break the rules](#style--rules).
+
+<a id="word-list--word-list"></a>
+
+#### Word list
+
+All word list entries have a link
+icon next to them. To link directly to an entry, you can right-click and
+copy the link address, or click and copy the URL from your address bar.
+
+Some word list entries include guidance to *avoid* or *don't use* a
+term. Apply this guidance as follows:
+
+- **Use with caution**. A recommendation to avoid using the term *when
+  possible*, or to use the term with caution. The term might be ambiguous
+  or obscure, so we provide alternative term suggestions or suggest that you
+  use a more specific term. However, you can use the term if needed. Where
+  appropriate, define the term or use it only once, as explained on the
+  [Jargon](#jargon) page.
+- **Don't use**. In all cases, we prefer to *not use the term*. The
+  term might be particularly ambiguous or it might have an offensive or
+  non-inclusive association. If such a term appears in code, we recommend that
+  you
+  [replace or write around the term](#inclusive-documentation--replace-or-write-around-non-inclusive-terms).
+- **Android**. Applies only to Android documentation.
+- **Google Cloud**. Applies only to Google Cloud documentation.
+- **Google Workspace**. Applies only to Google Workspace documentation.
+
+<a id="word-list--letter-number"></a>
+
+##### Numbers and Symbols
+
+<a id="word-list--u002b"></a>
+
+**+**
+    OK to use *+* with numbers in text, such as *customer records with
+    300+ demographic attributes*, except in formal contexts.
+<a id="word-list--ampersand"></a>
+
+**Avoid: & (ampersand)**
+    Don't use *&* instead of *and* in headings, text, navigation, or
+    tables of contents.
+    It's OK to use *&* when referencing UI elements that use *&*, or
+    in table headings and diagram labels where space constraints require
+    abbreviation.
+    It's OK to use `&` for technical purposes in code.
+<a id="word-list--2-step-verification"></a>
+
+**2-Step Verification**
+    When referring to Google's
+    [2-Step Verification](https://www.google.com/landing/2step/),
+    use initial caps.
+    When referring to
+    [generic 2-step verification](http://searchsecurity.techtarget.com/definition/two-step-verification),
+    use lowercase.
+
+<a id="word-list--letter-a"></a>
+
+##### A
+
+<a id="word-list--a-an"></a>
+
+**a and an**
+    Use *a* when the next word starts with a consonant *sound*,
+    regardless of what letter it starts with. For more information, see [Articles (a, an, the)](#articles).
+<a id="word-list--ab"></a>
+
+**A/B testing**
+    Capitalize and use slash notation for *A/B*.
+<a id="word-list--abnormal"></a>
+
+**Avoid: abnormal**
+    Don't use to refer to a person.
+    OK to use to refer to a condition of a computer system.
+<a id="word-list--abort"></a>
+
+**Avoid: abort**
+    Avoid in general usage. Instead, use words like *stop*, *exit*,
+    *cancel*, or *end*. In Linux, *abort* refers to a type of
+    signal that terminates an abnormal process.
+<a id="word-list--about-on"></a>
+
+**about versus on**
+    When a cross-reference includes information that describes what the
+    cross-reference links to, use *about* instead of *on*.
+    Recommended: For more information
+    about indexes, see [Managing indexes](https://cloud.google.com/firestore/docs/query-data/indexing).
+    Not recommended: For more information
+    on indexes, see [Managing
+    indexes](https://cloud.google.com/firestore/docs/query-data/indexing).
+<a id="word-list--above"></a>
+
+**Avoid: above**
+    Don't use for a range of version numbers. Instead, use
+    [*later*](#word-list--later).
+    Don't use to refer to a position in a document. Instead, use
+    *earlier* or *preceding*.
+    Don't use to refer to a position in the UI. Instead, write instructions
+    that avoid directional language. For more information,
+    see [Writing accessible documentation](#accessibility).
+    It's OK to use *above* in a non-directional way, such as when describing a hierarchy.
+<a id="word-list--access"></a>
+
+**Avoid: access (verb)**
+    Avoid when you can. Instead, use friendlier words like *see*,
+    *edit*, *find*, *use*, or *view*.
+<a id="word-list--access-token"></a>
+
+**access token**
+    Lowercase except at the beginning of a sentence,
+    heading, or list item.
+<a id="word-list--account-name"></a>
+
+**Don't use: account name**
+    Don't use. Instead, use [*username*](https://developers.google.com/style/word-list#username).
+<a id="word-list--actionable"></a>
+
+**Avoid: actionable**
+    Avoid unless it's the clearest and simplest phrasing for your audience.
+    Instead, leave it out or replace it with a phrase like *that you can act
+    on* or *useful*.
+    Don't use *actionable* in the legal sense without consulting a
+    lawyer.
+<a id="word-list--action-bar"></a>
+
+**Don't use: (Android) action bar**
+    In Android documentation, don't use. Instead, use
+    [*app bar*](#word-list--app-bar).
+<a id="word-list--ad-tech"></a>
+
+**ad tech**
+    Write out on first mention: *advertising technology (ad tech)*.
+    Don't use *adtech* or *ad-tech*.
+<a id="word-list--address-bar"></a>
+
+**address bar**
+    Use to refer to the URL bar or the combined URL bar and search box in a
+    browser.
+    Don't use *omnibox*.
+<a id="word-list--ad-hoc"></a>
+
+**ad hoc**
+    OK to use in database and analytics contexts to mean "free-form" or
+    "user-written" (for example, *ad hoc queries* or *an ad hoc
+    chart*). For other contexts, try to find a more specific English
+    equivalent.
+    Don't hyphenate or italicize the term.
+<a id="word-list--admin"></a>
+
+**Don't use: admin**
+    Write out *administrator* unless it's the name of a UI label or other
+    element.
+    It's OK to use *admin* in Android
+    documentation.
+<a id="word-list--administrator"></a>
+
+**Don't use: (Android) administrator**
+    In Android documentation, don't use. Instead, use *admin*.
+<a id="word-list--advertised-route-priority"></a>
+
+**advertised route priority**
+    OK to also use *base advertised route priority* when discussing
+    region-to-region costs.
+    Don't shorten or use variations of these terms.
+<a id="word-list--agnostic"></a>
+
+**Don't use: agnostic**
+    Don't use. Instead, use a more precise term like
+    *platform-independent*.
+<a id="word-list--ai"></a>
+
+**AI**
+    In general, you can use *AI* without spelling out *artificial intelligence*.
+    Most readers are familiar with the abbreviation *AI*. If you think your audience isn't
+    familiar with the term, spell it out on first use.
+<a id="word-list--aka"></a>
+
+**Don't use: aka**
+    Don't use. Instead, write out *also known as*, or present an
+    alternative term using parentheses or the word *or*. You can also
+    write out a definition.
+    Recommended:
+    Geographic data, also known as geospatial data, is ...
+    Recommended: Geographic data
+    (geospatial data) is ...
+    Recommended: Geographic data, or
+    geospatial data, is ...
+<a id="word-list--all-apps-screen"></a>
+
+**(Android) all apps screen**
+    In Android documentation: Lowercase except at the beginning of a sentence,
+    heading, or list item.
+<a id="word-list--allowlist"></a>
+
+**Don't use: allowlist (verb), allowlisted, allowlisting**
+    Don't use as a verb. Instead, rewrite to improve clarity.
+    OK to use *allowlist* as a noun.
+    For more information, see [blacklist](#word-list--blacklist).
+<a id="word-list--allows-you-to"></a>
+
+**Don't use: allows you to**
+    Don't use. Instead, use *lets you*. For more information, see [enable](#word-list--enable).
+<a id="word-list--alpha"></a>
+
+**alpha**
+    Lowercase except when part of a product name.
+    Recommended: PRODUCT_NAME
+    Alpha
+    Recommended: PRODUCT_NAME
+    is in alpha.
+<a id="word-list--america"></a>
+
+**Avoid: America, American**
+    Use only to refer to the *Americas* or the *American continent*.
+    Don't use to refer to the United States. Instead, use a more precise term
+    like *the US* or *the United States*, and *people in the
+    US*. For more information, see [US](#word-list--us).
+<a id="word-list--among"></a>
+
+**among**
+    See [between versus among](#word-list--between).
+<a id="word-list--am-pm"></a>
+
+**AM, PM**
+    To be consistent with [Material Design](https://material.io/design/communication/data-formats.html#date-and-time),
+    use all caps, no periods, and a space before.
+    Recommended: 9:00 AM
+    Recommended: 10:30 PM
+<a id="word-list--and-or"></a>
+
+**Avoid: and/or**
+    Don't use unless space is limited, such as in a table. For more
+    information, see [Slashes](#slashes--and-or).
+<a id="word-list--android"></a>
+
+**(Android) Android**
+    When referring to the operating system, capitalize *Android*.
+<a id="word-list--android-powered"></a>
+
+**(Android) Android-powered device**
+    Not *Android device*.
+<a id="word-list--and-so-on"></a>
+
+**Avoid: and so on**
+    Avoid using *and so on* whenever possible. For more information,
+    see [etc.](#word-list--etc)
+<a id="word-list--anti"></a>
+
+**anti***
+    See [guidance about hyphens with prefixes](#hyphens--prefixes).
+<a id="word-list--anti-pattern"></a>
+
+**Avoid: anti-pattern**
+    Avoid using *anti-pattern*, particularly as a standalone heading.
+    Instead, consider using a more specific and broadly understood term.
+    Recommended: Avoid these five SQL
+    errors.
+    Recommended: Avoid these five
+    programming practices that make SQL queries inefficient.
+    Not recommended: Avoid these five SQL
+    anti-patterns.
+<a id="word-list--api"></a>
+
+**API**
+    Use *API* to refer to either a web API or a language-specific API.
+    Don't use *API* when referring to a method or a class. For example,
+    don't write *This resource has one API* to mean "This resource has
+    one method."
+<a id="word-list--api-console"></a>
+
+**Don't use: API Console, APIs console, developer console, dev console, or Google API Console**
+    Don't use. Instead, refer to the *Google APIs Explorer* or to the
+    *Google Cloud console*. For more information, see
+    [console](#word-list--console).
+<a id="word-list--api-console-key"></a>
+
+**API Console key**
+    In most contexts, use *API key* instead of *API Console key*.
+    In Apps admin APIs, it's OK to use *API Console key* to distinguish
+    from other API keys.
+<a id="word-list--api-key"></a>
+
+**API key**
+    Not *developer key* or *dev key*.
+<a id="word-list--apis-explorer"></a>
+
+**APIs Explorer**
+    Not *API explorer* or other variants.
+<a id="word-list--app"></a>
+
+**app**
+    In general, use *app* instead of *application* when referring to
+    programs for end users, especially in the context of mobile or web
+    software.
+    In some contexts, such as enterprise software, it's OK to use
+    *application* to convey a sense of greater complexity.
+    Use *application* in standard phrases such as *application
+    programming interface*.
+<a id="word-list--app-bar"></a>
+
+**(Android) app bar**
+    In Android contexts, formerly *action bar*.
+<a id="word-list--appendix"></a>
+
+**appendix**
+    Use the plural *appendixes*, not *appendices*.
+<a id="word-list--application"></a>
+
+**application**
+    See [app](#word-list--app).
+<a id="word-list--as"></a>
+
+**as**
+    If you mean *because*, then use *because* instead of
+    *as*. *As* is ambiguous; it can refer to the passage of time.
+    *Because* refers to causation or the reason for something.
+<a id="word-list--as-of-this-writing"></a>
+
+**Avoid: as of this writing**
+    Avoid because this phrase is implied. The phrase can also prematurely
+    disclose product or feature strategy or inappropriately imply that a
+    product or feature might change.
+    See also [currently](#word-list--currently) and [presently](#word-list--presently).
+    Recommended: BigQuery doesn't support
+    that function.
+    Not recommended: As of this writing,
+    BigQuery doesn't support that function.
+    For more information, see [Timeless
+    documentation](#timeless-documentation).
+<a id="word-list--authentication-and-authorization"></a>
+
+**authentication and authorization**
+    In general, use the word *authenticated* only to refer to users,
+    and use *authorized* only to refer to requests that are sent by a
+    client app on behalf of an authenticated user.
+
+    A user *authenticates* their identity by entering their password
+    (or giving some other proof of identity). The *authenticated
+    user* then *authorizes* the client app to send an
+    *authorized request* to the server on the user's behalf.
+    When you want to use a preposition with *authenticate*, use
+    *against*.
+<a id="word-list--authn-authz"></a>
+
+**Don't use: authN, authZ**
+    Don't use. Instead, use *authentication* or *authorization*.
+<a id="word-list--auto"></a>
+
+**auto***
+    See [guidance about hyphens with prefixes](#hyphens--prefixes).
+<a id="word-list--autohealing"></a>
+
+**autohealing**
+    Not *auto-healing*.
+<a id="word-list--auto-mode-vpc"></a>
+
+**auto mode VPC network**
+    Not *auto mode network*.
+<a id="word-list--autopopulate"></a>
+
+**autopopulate**
+    Not *auto populate* or *auto-populate*.
+<a id="word-list--autoscaling"></a>
+
+**autoscaling**
+    Not *auto-scaling*.
+<a id="word-list--autotagging"></a>
+
+**autotagging**
+    Not *auto-tagging*.
+<a id="word-list--autoupdate"></a>
+
+**Don't use: autoupdate**
+    Don't use. Instead, use *automatically update*.
+<a id="word-list--aware"></a>
+
+**Avoid: -aware**
+    Avoid using as a compound modifier, as in *healthcare-aware*.
+    OK to use when it's part of a product name, such as *Identity-Aware
+    Proxy*.
+
+<a id="word-list--letter-b"></a>
+
+##### B
+
+<a id="word-list--backend"></a>
+
+**backend**
+    Not *back-end* or *back end*.
+<a id="word-list--bar"></a>
+
+**Avoid: bar**
+    Avoid when possible. For more information, see [foo](#word-list--foo).
+<a id="word-list--bare-metal"></a>
+
+**bare metal**
+    Lowercase except at the beginning of a sentence,
+    heading, or list item.
+    Hyphenate when used as a compound modifier, such as *bare-metal
+    server*.
+<a id="word-list--base64"></a>
+
+**base64**
+    Lowercase except at the beginning of a sentence,
+    heading, or list item. Otherwise, capitalize *Base64* only if it's part of a
+    formal name.
+    Write *base64* in code font *only* if it's a string literal or
+    otherwise quoted from code.
+<a id="word-list--baz"></a>
+
+**Avoid: baz**
+    Avoid when possible. For more information, see [foo](#word-list--foo).
+<a id="word-list--below"></a>
+
+**Avoid: below**
+    Don't use for a range of version numbers. Instead, use
+    [*earlier*](#word-list--earlier).
+    Don't use to refer to a position in a document. Instead, use *later*
+    or *following*.
+    Don't use to refer to a position in the UI. Instead, write instructions
+    that avoid directional language. For more information, see
+    [Writing accessible documentation](#accessibility).
+    It's OK to use *below* in set phrases such as *below (the)
+    average*, *below the mean*, *below zero*.
+    It's OK to use *below* in a non-directional way, such as when describing a hierarchy.
+<a id="word-list--best-effort"></a>
+
+**Avoid: best effort**
+    Avoid where possible. Instead, use more specific wording. After providing
+    a description, you can add a phrase like "sometimes referred to as *best
+    effort*."
+<a id="word-list--beta"></a>
+
+**beta**
+    Lowercase except when part of a product name.
+    Recommended: PRODUCT_NAME
+    Beta
+    Recommended: PRODUCT_NAME
+    is currently in beta.
+<a id="word-list--between"></a>
+
+**between versus among**
+    It's fine to use *between* when talking about more than two things;
+    however, *between* isn't interchangeable with *among*.
+    Use *between* when you're talking about two or more distinct
+    things:
+    Recommended: JavaScript introduces
+    dependencies between the DOM, the CSSOM, and JavaScript execution.
+    Use *among* when you're talking about things that are part of a group
+    or things that aren't distinct:
+    Recommended: ... a conventional SQL
+    database that can be shared among multiple apps.
+    More examples:
+    Recommended: Because screen
+    dimensions vary widely among devices (for example, between phones and
+    tablets, and even among different phones), you should configure the
+    viewport so that your pages render correctly on many different devices.
+    Not recommended: Because screen
+    dimensions vary widely between devices (for example, between phones and
+    tablets, and even between different phones), you should configure the
+    viewport so that your pages render correctly on many different devices.
+    Recommended: You can share services
+    among multiple clients.
+    Not recommended: You can share
+    services between multiple clients.
+    See also [Grammar Girl's
+    discussion of *between* and *among*](http://www.quickanddirtytips.com/education/grammar/between-versus-among).
+<a id="word-list--big-endian"></a>
+
+**big-endian**
+    Hyphenate. Lowercase except at the beginning of a sentence,
+    heading, or list item.
+    Recommended: The codebase assumes
+    big-endian byte ordering.
+    Not recommended: The codebase assumes
+    Big Endian byte ordering.
+    Not recommended: The codebase assumes
+    Big-endian byte ordering.
+    Not recommended: The codebase assumes big
+    endian byte ordering.
+<a id="word-list--billing-charges"></a>
+
+**Avoid: billing charges**
+    Don't use *billing charges* to mean charges that appear on a bill.
+    Instead, use *billed charges*.
+    Use *billing charges* to describe the cost of creating the bill.
+<a id="word-list--black-box"></a>
+
+**Avoid: black-box**
+    Avoid using *black-box*, *blackbox*, or *black box* to
+    describe monitoring and testing. Consider using a more precise term for
+    clarity.
+
+    - For monitoring, use *synthetic monitoring*.
+    - For testing, use *opaque-box testing*.
+<a id="word-list--black-friday"></a>
+
+**Avoid: Black Friday**
+    Avoid unless explicitly referring to an event in the US. Instead use
+    *peak scale event*.
+<a id="word-list--blackhat"></a>
+
+**Don't use: blackhat, black hat, black-hat**
+    Don't use. Instead, use precise terms for the kind of violation or
+    practice, such as *illegal*, *unethical*, or *in violation of
+    rules*.
+<a id="word-list--blackhole"></a>
+
+**Don't use: blackhole (verb), blackholed (adjective)**
+    Don't use. Instead, use a more descriptive term or phrase, such as
+    *dropped without notification*.
+<a id="word-list--blacklist"></a>
+
+**Don't use: blacklist, black list, black-list**
+    Don't use *blacklist*, *whitelist*, and *graylist*.
+    Instead, use more precise terms that are appropriate for your domain.
+    - For the noun *blacklist*, consider using a replacement such as
+      *denylist*, *excludelist*, or *blocklist*.
+    - For the noun *whitelist*, consider using a replacement such as
+      *allowlist*, *trustlist*, or *safelist*.
+    - For the noun *graylist* (*greylist*), consider using a
+      replacement such as *provisional list*.
+    In all of these cases, consider that there might not actually be a list
+    involved. When replacing problematic terms, be sure to be technically
+    accurate for the specific context.
+    For the verb forms of these words, a simple word-for-word replacement
+    typically isn't the best solution. Instead, replace verbs such as
+    *blacklisted* with phrases that accurately convey the relevant
+    action. For example:
+    Recommended: To deny requests from
+    an IP address, add it to the `dos.yaml` file.
+    Not recommended: To denylist an IP
+    address, add it to the `dos.yaml` file.
+    Don't use: To blacklist an IP
+    address, add it to the `dos.yaml` file.
+    If the command or code that you're documenting uses one of these words,
+    then use the words only in direct reference to the code items
+    ([formatted as code](#code-in-text)), and make it clear
+    what you're referring to.
+    Recommended: Add a user to the
+    allowlist (`whitelist`) by entering the following:
+    `whitelist adduser EMAIL_ADDRESS`.
+    Not recommended: Add a user to the
+    whitelist by entering the following: `whitelist adduser
+    EMAIL_ADDRESS`.
+    For more information, see the
+    [inclusive documentation](#inclusive-documentation) page.
+<a id="word-list--blacklisted"></a>
+
+**Don't use: blacklisted, black listed, black-listed**
+    Don't use. See [blacklist](#word-list--blacklist).
+<a id="word-list--blacklisting"></a>
+
+**Don't use: blacklisting, black listing, black-listing**
+    Don't use. See [blacklist](#word-list--blacklist).
+<a id="word-list--blast-radius"></a>
+
+**Don't use: blast radius**
+    Don't use. Instead, use a more precise term like *affected area* or
+    *spatial impact*.
+<a id="word-list--blind"></a>
+
+**Avoid: blind**
+    Avoid using *blind to* or *blind eye to*. Instead, use more
+    precise terms like *ignore*, *unaware of*, *disregard*,
+    *avoid*, or *reject*.
+    Avoid using *blind writes*. Instead, use a more precise phrase, such
+    as *a write operation without a read operation*.
+    Avoid using *blind change* or *change blindly*. Instead, use a
+    more precise phrase such as *change without first confirming the
+    value*.
+    When referring to people, use terms like *person who is blind*,
+    *screen reader user* (if applicable), *person who is visually
+    impaired*, *person who is low-vision*, *magnification user*
+    (if applicable).
+<a id="word-list--blue-green"></a>
+
+**blue-green**
+    Not *blue/green* or *blue green*.
+<a id="word-list--boolean"></a>
+
+**boolean**
+    In most contexts, *boolean* refers to a specific data type in a
+    specific programming language. In such cases, use code font and the exact
+    spelling and capitalization of the programming keyword.
+    When referring to the abstract data type, use lowercase.
+    If you refer to *Boolean mathematics* or *Boolean logic*, use
+    uppercase.
+<a id="word-list--branding-information"></a>
+
+**branding information**
+    In the Google Cloud console, the phrase *branding information* refers
+    to the information that Google shows to users when the client asks them to
+    authorize access: specifically, the project's name and logo, and the
+    developer's Google Account. This information is set in the **Consent
+    screen** page.
+<a id="word-list--break-glass"></a>
+
+**Don't use: break-glass**
+    Don't use. Instead, use a more precise term depending on context:
+
+    - To describe a general emergency or procedure that grants emergency
+      access, use *emergency access*.
+    - To describe a fallback procedure, use *manual fallback* or
+      *preplanned procedure*.
+<a id="word-list--brown-bag"></a>
+
+**Don't use: brown bag, brown-bag**
+    Don't use. Instead, use a more precise term like *learning session*,
+    *lunch and learn*, *lunchtime learning session*,
+    *casual training*, or *informal training*.
+<a id="word-list--build-cop"></a>
+
+**Don't use: build cop, build sheriff**
+    Don't use. Instead, use a more precise term like *build monitor*.
+<a id="word-list--button"></a>
+
+**button**
+    In a UI, a link isn't the same as a button; don't use the term
+    *button* to refer to a link.
+    Use *button* to refer to mechanical buttons (like the volume control
+    buttons on the side of a phone) and capacitive touch buttons on a phone
+    (like the Home button). You *press* mechanical buttons, and
+    *tap* capacitive and on-screen buttons.
+
+<a id="word-list--letter-c"></a>
+
+##### C
+
+<a id="word-list--can"></a>
+
+**can**
+    Use *can* in the following ways:
+
+    - To convey permission or ability (for example, "You can access the
+      server").
+    - To refer to an optional action (for example, "You can also view
+      logs with the Log Viewer").
+    - To describe a possible outcome (for example, "The process can
+      take 30 minutes").
+    See also [could](#word-list--could), [may](#word-list--may),
+    [might](#word-list--might), [must](#word-list--must),
+    [should](#word-list--should), and [would](#word-list--would).
+    For information about clarifying who's performing an action, see
+    [Active voice](#voice).
+<a id="word-list--canary"></a>
+
+**Avoid: canary**
+    Don't use *canary* as a verb, and don't use *canarying*.
+    When possible, avoid [jargon](#jargon) like *canary* and
+    *canary testing*. If you use one of these phrases, define it on first
+    use or provide a link to the definition, and use it consistently
+    throughout the document.
+<a id="word-list--cell-phone"></a>
+
+**Don't use: cell phone, cellphone**
+    Don't use. Instead, use *mobile phone*, or if you're talking about
+    more than phones, then use *mobile device*.
+    It's OK to use *phone* (without *mobile*) when the context is
+    clear.
+<a id="word-list--cellular-data"></a>
+
+**Don't use: cellular data**
+    Don't use. Instead, use *mobile data*.
+<a id="word-list--cellular-network"></a>
+
+**Don't use: cellular network**
+    Don't use. Instead, use *mobile network*.
+<a id="word-list--chapter"></a>
+
+**chapter**
+    When referring to documentation that isn't in the form of a book, don't
+    use the term *chapter*. Instead, refer to documents, pages, or
+    sections.
+<a id="word-list--check"></a>
+
+**Avoid: check**
+    Don't use to refer to marking a checkbox. Instead, use *select*.
+    Recommended: Select **Automatically
+    check for updates**.
+    Not recommended: Check **Automatically
+    check for updates**.
+<a id="word-list--checkbox"></a>
+
+**checkbox**
+    Not *check box*.
+<a id="word-list--choose"></a>
+
+**choose**
+    *Choose* is fine to use for generic contexts. For UI elements, use
+    [select](#word-list--select).
+<a id="word-list--chubby"></a>
+
+**Don't use: chubby**
+    Don't use. Instead, use a word that clearly explains what you mean, such
+    as *unused* or *overextended*.
+<a id="word-list--clear"></a>
+
+**clear**
+    Use (as a verb) to refer to clearing a check mark from a checkbox.
+    Recommended: Clear **Automatically
+    check for updates**.
+    Not recommended: Uncheck
+    **Automatically check for updates**.
+    Not recommended: Deselect
+    **Automatically check for updates**.
+<a id="word-list--cli"></a>
+
+**Avoid: CLI**
+    Don't use *CLI* generically to refer to a command-line interface.
+    Instead, refer to the specific command-line interface, such as the
+    [Google Cloud CLI](#word-list--gcloud).
+<a id="word-list--click"></a>
+
+**click**
+    When the environment is a desktop with a mouse, use *click* for most
+    targets, such as buttons, links, list items, and radio buttons. Don't use
+    *click on*.
+    Recommended: Click **OK**.
+    Not recommended: Click on **OK**.
+    Hyphenate *right-click*, *left-click*, and *double-click*.
+    When a click or tap action reveals a collapsed list, you can write
+    *click to expand* or simply *expand*.
+    It's OK to write *click in* when referring to a region that needs
+    focus (for example: *click in the window*), but not when referring to
+    a control or a link.
+    For Android apps, don't use
+    *click*. Instead, use [tap](#word-list--tap).
+<a id="word-list--click-here"></a>
+
+**Don't use: click here**
+    Don't use. For information and alternatives, see
+    [Avoid vague link text](#cross-references--vague-link-text).
+<a id="word-list--clickthrough"></a>
+
+**clickthrough (noun), click through (verb)**
+<a id="word-list--client"></a>
+
+**client**
+    In REST and RPC API documentation, *client* is short for *client
+    app*—that is, the app that the developer is writing.
+    Don't use *client* as an abbreviation for *client library*;
+    instead, use *library*.
+<a id="word-list--client-id"></a>
+
+**client ID**
+    Lowercase except at the beginning of a sentence,
+    heading, or list item.
+<a id="word-list--client-secret"></a>
+
+**client secret**
+    Lowercase except at the beginning of a sentence,
+    heading, or list item.
+<a id="word-list--cloud"></a>
+
+**Avoid: (Cloud) Cloud**
+    Don't use as short for *Google Cloud*.
+    For generic references such as *the cloud* or *hybrid cloud*,
+    use lowercase.
+<a id="word-list--gcp-console"></a>
+
+**Don't use: (Cloud) Cloud console**
+    Don't use. Instead, refer to the full name *Google Cloud console*.
+    If you aren't discussing any other console (such as the Google Admin
+    console), you can abbreviate to *the console* after first mention.
+    Use *the* before the tool name. For more information, see
+    [console](#word-list--console).
+<a id="word-list--cloud-sdk"></a>
+
+**Cloud SDK**
+    Not *Google Cloud SDK*.
+<a id="word-list--co"></a>
+
+**co***
+    See [guidance about hyphens with prefixes](#hyphens--prefixes).
+<a id="word-list--codebase"></a>
+
+**codebase**
+    Not *code base*.
+<a id="word-list--codelab"></a>
+
+**codelab**
+    Not *code lab* or *code-lab*. For more information, see
+    [documentation](#word-list--documentation).
+<a id="word-list--cold"></a>
+
+**cold**
+    When possible, avoid [jargon](#jargon) like *cold
+    failover*, *cold standby*, and *cold spare*. If you use one
+    of these phrases, define it on first use and use it consistently
+    throughout the document.
+<a id="word-list--colocate"></a>
+
+**colocate**
+    Not *co-locate* or *colo*.
+<a id="word-list--compliant"></a>
+
+**Avoid: compliant, compliance**
+    Use with caution. A claim that a product or its output is *compliant*
+    with a standard is a strong statement.
+<a id="word-list--comprise"></a>
+
+**Don't use: comprise**
+    Don't use. Instead, use *consist of*, *contain*, or
+    *include*.
+<a id="word-list--config"></a>
+
+**Avoid: config**
+    Avoid when possible. Instead, spell out the full word when it's used in a
+    non-code sense: *configuration* or *configuring*. Use the
+    verbatim code item name when referring to, for example, a data structure
+    or a file with that name.
+<a id="word-list--confidential"></a>
+
+**confidential**
+    *Confidential* data is data that is protected to prevent unauthorized access. See
+    [sensitive](#word-list--sensitive).
+<a id="word-list--cons"></a>
+
+**Don't use: cons**
+    Don't use. Instead, use a more precise term, such as *disadvantages*.
+<a id="word-list--console"></a>
+
+**Avoid: console**
+    Don't use in isolation. Instead, use the name of the specific console,
+    such as the [Google Cloud
+    console](https://console.cloud.google.com/) or the Google Admin console.
+    Use *the* before the name of a console.
+    After giving the full name of a console, you can use a shortened version
+    of the name, such as the *Admin console*.
+    If you're only discussing the Google Cloud console, after giving the full
+    name you can refer to *the console*.
+    To refer to a sub-page of a console, use the term *page*.
+    If a specific term for a browser-based interface is unavailable, use
+    *web interface*.
+<a id="word-list--content-type"></a>
+
+**Avoid: content type**
+    Be as specific as possible when writing about a content type, and use the term only when applicable.
+    For example, you can use this term if you're referring to the value of the `Content-Type` HTTP header.
+    Also see [media type](#word-list--media-type).
+<a id="word-list--control-keys"></a>
+
+**Control+S, Command+S, and other keyboard commands**
+    To refer to a `Control` character, use
+    `Control`+CHARACTER.
+    Don't use *Ctl-S*, *Cmd-S*, or *Cloverleaf-S*.
+    In most cases, use an uppercase letter for CHARACTER.
+    In macOS, many keyboard commands use the `Command` key instead of
+    the `Control` key, and there's an `Option` key instead
+    of an `Alt` key. If your audience includes macOS users and
+    Windows or Linux users, then mention both keyboard commands.
+    Recommended: `Control+S`
+    (`Command+S` on macOS)
+<a id="word-list--copy-paste"></a>
+
+**Avoid: Copy and paste**
+    Avoid using. Instead, explain what to enter into a field and not how.
+    Recommended: In the
+    **Query** field, enter the output from the previous step.
+    Not recommended: Copy the output from
+    the previous step and paste into the **Query** field.
+<a id="word-list--could"></a>
+
+**Avoid: could**
+    Avoid using. Instead, use *can* where possible.
+    See also [can](#word-list--can), [may](#word-list--may),
+    [might](#word-list--might), [must](#word-list--must),
+    [should](#word-list--should) and [would](#word-list--would).
+    For information about clarifying who's performing an action, see
+    [Active voice](#voice).
+    For information about tenses, see [Present
+    tense](#tense).
+<a id="word-list--cpu"></a>
+
+**CPU**
+    All caps. No need to expand the abbreviation on first mention.
+<a id="word-list--crazy"></a>
+
+**Don't use: crazy, bonkers, mad, lunatic, insane, loony**
+    Don't use. Instead, use *complicated*, *complex*,
+    *baffling*, *strange*, or *unexpected*, and only for
+    inanimate objects.
+<a id="word-list--create-new"></a>
+
+**Avoid: Create a new ...**
+    Avoid using unless you need to distinguish the item from another recently
+    created item. Instead, use *Create a ...*
+    Recommended: Create a project.
+    Not recommended: Create a new project.
+<a id="word-list--cripple"></a>
+
+**Don't use: cripple**
+    Don't use. Instead, use more precise language. For example, instead of
+    *it crippled the server*, write *it slowed the server down*.
+    When referring to people, use terms that specifically describe a physical
+    impairment, such as *person with a motor disability*; *person with
+    a mobility impairment* (refers to walking or moving about); *person
+    with dexterity impairment* (refers to using a standard mouse or
+    keyboard); *person who uses a wheelchair, walker, or cane*;
+    *wheelchair user*; *person with restricted or limited mobility*.
+<a id="word-list--cross-site-request-forgery"></a>
+
+**cross-site request forgery**
+    Lowercase except at the beginning of a sentence,
+    heading, or list item.
+<a id="word-list--curated-roles"></a>
+
+**Don't use: curated roles**
+    Don't use. Instead, use *predefined roles*.
+<a id="word-list--currently"></a>
+
+**Avoid: currently**
+    Avoid because this word is implied. The word can also prematurely disclose
+    product or feature strategy or inappropriately imply that a product or
+    feature might change.
+    See also
+    [as of this writing](#word-list--as-of-this-writing) and
+    [presently](#word-list--presently).
+    Recommended: Windows isn't supported.
+    Not recommended: Windows isn't
+    currently supported.
+    For more information, see
+    [Timeless documentation](#timeless-documentation).
+<a id="word-list--custom-mode-vpc-network"></a>
+
+**custom mode VPC network**
+    Not *custom mode network*.
+<a id="word-list--curl"></a>
+
+**curl**
+    Not *cURL*.
+    For information about when to use code format, see
+    [Items that are sometimes in code font](#code-in-text--items-that-are-sometimes-in-code-font).
+<a id="word-list--cyber-monday"></a>
+
+**Cyber Monday**
+    Avoid unless explicitly referring to an event in the US. Instead use
+    *peak scale event*.
+
+<a id="word-list--letter-d"></a>
+
+##### D
+
+<a id="word-list--dash"></a>
+
+**dash**
+    A dash (`—`) isn't the same character as a hyphen
+    (`-`). The characters are used for different purposes.
+    Therefore, don't use the word *dash* to refer to a hyphen.
+<a id="word-list--dashboard"></a>
+
+**Avoid: dashboard**
+    Don't use to refer to the Google Cloud console. For more information, see
+    [console](#word-list--console).
+    Use *dashboard* not *Dashboard* unless it's officially part of a
+    product name.
+<a id="word-list--data"></a>
+
+**data**
+    Use *data* as singular, not plural; *the data is*, not
+    *the data are*.
+    Use data as a mass noun, not a count noun; *less data*, not
+    *fewer data*.
+<a id="word-list--data-center"></a>
+
+**data center**
+    Not *datacenter*.
+<a id="word-list--data-center-campus"></a>
+
+**data center campus**
+    Use when referring to an entire physical location, which can encompass one
+    or more data centers.
+<a id="word-list--data-cleaning"></a>
+
+**data cleaning**
+    Not *data cleansing*.
+<a id="word-list--dataflow"></a>
+
+**data flow (noun); dataflow (noun)**
+    If it's possible to replace with the phrase *flow of data*, then use
+    two words: *data flow*.
+    If that replacement doesn't work, such as when referring to something like
+    stream processing or reactive programming, then use one word:
+    *dataflow*.
+<a id="word-list--data-source"></a>
+
+**data source**
+    Not *datasource*.
+<a id="word-list--datastore"></a>
+
+**datastore**
+    Not *data store*.
+<a id="word-list--data-type"></a>
+
+**data type**
+    Not *datatype*.
+<a id="word-list--dead-letter"></a>
+
+**dead-letter queue, dead letter**
+    Define on first use, for example *dead-letter queue (unprocessed
+    messages queue)*.
+<a id="word-list--deep-linking"></a>
+
+**deep linking**
+    Not *deep-linking*. However, if you can replace with
+    *linking*, then do so.
+<a id="word-list--deficient"></a>
+
+**Avoid: deficient**
+    Don't use to refer to a person.
+    OK to use to refer to a condition of a computer system.
+<a id="word-list--deformed"></a>
+
+**Avoid: deformed**
+    Don't use to refer to a person.
+    OK to use to refer to a condition of a computer system or
+    inanimate object.
+<a id="word-list--dmz"></a>
+
+**Don't use: demilitarized zone (DMZ)**
+    Don't use. Instead, use a more precise term like *perimeter network*.
+<a id="word-list--denigrate"></a>
+
+**Don't use: denigrate**
+    Don't use. Instead, use *disparage*.
+<a id="word-list--denylisted"></a>
+
+**Don't use: denylist (verb), denylisted, denylisting**
+    Don't use as a verb. Instead, rewrite to improve clarity.
+    OK to use *denylist* as a noun.
+    For more information, see [blacklist](#word-list--blacklist).
+<a id="word-list--deprecate"></a>
+
+**deprecate**
+    To *deprecate* an item is to recommend against the item's use,
+    typically as a warning that the item will soon be unavailable or
+    unsupported. Don't use *deprecated* to mean *removed*,
+    *deleted*, *shut down*, or *turned down*.
+<a id="word-list--deselect"></a>
+
+**Don't use: deselect**
+    Don't use to refer to clearing a check mark from a checkbox. Instead, use
+    *clear*.
+    Recommended: Clear **Automatically
+    check for updates**.
+    Not recommended: Deselect
+    **Automatically check for updates**.
+    Not recommended: Uncheck
+    **Automatically check for updates**.
+<a id="word-list--desire"></a>
+
+**Don't use: desire, desired**
+    Don't use. Instead, use a word like *want* or *need*.
+    Recommended: Set the value to the
+    size that you want.
+    Not recommended: Set the value to
+    the size that you desire.
+    Not recommended: Set the value to
+    the desired size.
+<a id="word-list--developers-console"></a>
+
+**Don't use: Developers Console**
+    Don't use. For more information, see [console](#word-list--console).
+<a id="word-list--devops"></a>
+
+**DevOps**
+    Short for *development operations*. No need to spell out on first
+    mention unless the audience requires it. For more information, see [DevOps](https://wikipedia.org/wiki/DevOps).
+<a id="word-list--dialog"></a>
+
+**dialog**
+    Use *dialog* for the UI element sometimes called a [dialog box](http://wikipedia.org/wiki/Dialog_box).
+    Use *dialogue* only for verbal interaction between people.
+<a id="word-list--directory"></a>
+
+**directory, folder**
+    If the context that you're documenting (such as an IDE's GUI) uses one
+    term or the other, use that term. If not, then use *directory* in a
+    command-line context, and *folder* in a GUI context. When in doubt,
+    default to *directory*.
+<a id="word-list--disable"></a>
+
+**Avoid: disable**
+    Don't use *disable* or *disabled* to describe something that's
+    broken.
+    When describing a user action or the state of a UI element, use a more
+    precise term where possible. You can use *inactive*,
+    *unavailable*, *deactivate*, *turn off*, or
+    *deselect*, depending on the context. Use the same term consistently throughout your
+    document.
+    See also [enable](#word-list--enable).
+<a id="word-list--disclosure-triangle"></a>
+
+**Don't use: disclosure triangle, disclosure widget**
+    Don't use. Instead, use *expander arrow*.
+<a id="word-list--display"></a>
+
+**display (verb)**
+    Don't use as an intransitive verb. *Display* is a transitive verb;
+    therefore, it requires an object. It is often misused in technical
+    documentation, as demonstrated by the following example:
+    Recommended: The Output Directories
+    area appears.
+    Recommended: The Output Directories
+    area is displayed.
+    Not recommended: The Output
+    Directories area displays.
+    The following example demonstrates correct usage of the verb
+    *display* but means something quite different from the preceding
+    examples.
+    Recommended: The Output Directories
+    area displays the vector image.
+<a id="word-list--ddos"></a>
+
+**distributed denial-of-service (DDoS)**
+    Hyphenate as shown. On subsequent mention, use *DDoS*.
+<a id="word-list--dns-server-policy"></a>
+
+**DNS server policy**
+    Lowercase *server policy*.
+<a id="word-list--dnskey"></a>
+
+**DNSKEY**
+    One word, all capital letters.
+<a id="word-list--documentation"></a>
+
+**documentation or document or documents**
+    To refer specifically to the text on a page that explains a product, feature, or service,
+    use *this document*, and not *this article*, *this topic*, *this doc*, or
+    *this page*. It's OK to use *this tutorial*, *this quickstart*, or *this
+    codelab* for those specific documentation types.
+
+    Always spell out *documentation* except in cases where space is limited, such as in
+    tabs and URLs.
+
+    See also
+    [page](#word-list--page).
+    Recommended: You can find many
+    examples in this document.
+    Not recommended: You can find many
+    examples in this article.
+    Recommended: This document provides
+    guidance about creating tables.
+    Not recommended: This page provides
+    guidance about creating tables.
+<a id="word-list--docset"></a>
+
+**documentation set**
+    Not *doc set* or *docset*.
+<a id="word-list--does-not-yet"></a>
+
+**Avoid: does not yet**
+    Avoid in timeless documentation because this phrase can become outdated.
+    The phrase can also prematurely disclose product or feature strategy or
+    inappropriately imply that a product or feature might change.
+    Recommended: The Google Cloud console
+    doesn't support this IAM role.
+    Not recommended: The Google Cloud
+    console does not yet support this IAM role.
+    For more information, see
+    [Timeless documentation](#timeless-documentation).
+<a id="word-list--dojo"></a>
+
+**Don't use: dojo**
+    Don't use. Instead, use a precise term that is accurate for the context,
+    such as *training* or *workshop*.
+<a id="word-list--domain-name-registrar"></a>
+
+**domain name registrar**
+    Lowercase except at the beginning of a sentence,
+    heading, or list item.
+<a id="word-list--dnssec"></a>
+
+**Domain Name System Security Extensions (DNSSEC)**
+    Write out and capitalize each word on first use. OK to abbreviate as
+    *DNSSEC* after first use.
+<a id="word-list--double-tap"></a>
+
+**(Android) double-tap**
+    Hyphenate. Lowercase except at the beginning of a sentence,
+    heading, or list item.
+<a id="word-list--downscope"></a>
+
+**Avoid: downscope**
+    Consider using a more descriptive term like *constrain scope* or
+    *reduce scope*. Because *downscope* might not be broadly
+    understood, if you use the term, make sure to define it on first use.
+        Don't use *down scope* or *down-scope*
+            Recommended: Reducing the scope of a
+            token helps you follow the principle of least privilege.
+            Recommended (first use): The IAM
+            recommender helps you *downscope* (reduce) the permissions that are
+            available to your users.
+<a id="word-list--drag"></a>
+
+**drag**
+    Use *drag*, not *click and drag* and not *drag and drop*.
+    OK to use *drag-and-drop* as an adjective.
+    Recommended: Drag the USER
+    to the **Authorized** box.
+<a id="word-list--drop-down"></a>
+
+**Avoid: drop-down**
+    In most cases, you can omit *drop-down* from phrases like *drop-down list* or
+    *drop-down menu*, and just use *list* or *menu*. Include *drop-down* as a
+    modifier only if the omission would cause ambiguity. Don't use *drop-down* as a
+    standalone noun.
+<a id="word-list--dumb-down"></a>
+
+**Don't use: dumb down**
+    Don't use. Instead, use a word or phrase what's happening, such as
+    *simplify* or *remove technical jargon*.
+<a id="word-list--dummy-variable"></a>
+
+**Don't use: dummy variable**
+    Don't use to refer to placeholders. Instead, use *placeholder*.
+    Also don't use if referring to the concept in statistics known as a
+    [dummy variable](https://en.wikipedia.org/wiki/Dummy_variable_(statistics)).
+    Instead, use alternate terms such as
+    *indicator variable*, *design variable*, *one-hot
+    encoding*, *Boolean indicator*, *binary variable*, or
+    *qualitative variable*.
+
+<a id="word-list--letter-e"></a>
+
+##### E
+
+<a id="word-list--each"></a>
+
+**each**
+    *Each* refers to every individual item taken individually, not to a
+    group of items taken collectively. In other words, *each* isn't a
+    synonym for *all*. For example, *a list of each item* is
+    ambiguous; *a list of all the items* or *a list of the items* is
+    generally clearer.
+<a id="word-list--earlier"></a>
+
+**earlier**
+    Use for a range of version numbers, not *lower*.
+    Recommended: Use version 2.2 or
+    earlier.
+    Not recommended: Use version 2.2 or
+    lower.
+    In Android documentation, don't use
+    *earlier* for a range of version numbers. Instead, use *lower*.
+    When referring to a position in a document, use *earlier* or
+    *preceding*, not *higher*.
+<a id="word-list--easy"></a>
+
+**Avoid: easy, easily**
+    What might be easy for you might not be easy for others. Try eliminating
+    this word from the sentence because usually the same meaning can be conveyed
+    without it.
+<a id="word-list--ecommerce"></a>
+
+**ecommerce**
+    Not *e-commerce*.
+<a id="word-list--edge-availability-domain"></a>
+
+**edge availability domain**
+    Don't use *edge availability zone*, *metro availability domain*,
+    or *metro availability zone*. Don't shorten to *EAD*.
+<a id="word-list--eg"></a>
+
+**Don't use: e.g.**
+    Don't use. Instead, use phrases like *for example* or *such as*.
+    Many people confuse *e.g.* and *i.e.*
+<a id="word-list--egress"></a>
+
+**egress**
+    When referring to the networking term, use lowercase.
+<a id="word-list--either"></a>
+
+**either**
+    When using *either*, use parallel syntax.
+    Recommended: Do either option 1 or
+    option 2.
+    Recommended: Either do option 1 or
+    do option 2.
+    Not recommended: Either do option 1
+    or option 2.
+    In general, use *either* only for a choice between two things, not
+    for a choice among multiple things. Writing *either A or B or C* will
+    distract some readers, but if it's the best phrasing for your situation,
+    then use it.
+<a id="word-list--element"></a>
+
+**element**
+    In HTML and XML, a tag is a component of an element that indicates
+    the start or end of the element. (For example, the
+    `<i>` start tag indicates the beginning of the
+    `<i>example</i>` element.) In general, don't use
+    the term *tag* to refer to an entire element.
+<a id="word-list--email"></a>
+
+**email**
+    Not *e-mail*, *Email*, or *E-mail*.
+    Don't use as a verb.
+    Use a specific verb in front of the word. For example, *send email*.
+    This construction is better for translation and a
+    [global audience](#translation).
+<a id="word-list--emoji"></a>
+
+**emoji**
+    Use *emoji* for both singular and plural forms. See [Don't
+    know the difference between emoji and emoticons? Let me explain](https://www.theguardian.com/technology/2015/feb/06/difference-between-emoji-and-emoticons-explained) and [What's the Plural of Emoji?](http://www.theatlantic.com/technology/archive/2016/01/whats-the-plural-of-emoji-emojis/422763/)
+<a id="word-list--enable"></a>
+
+**enable**
+    In procedures, use the appropriate label and action for the
+    [UI element](#ui-elements) that the user interacts with. When describing a
+    user action or the state of a UI element, use a more precise term where possible. It's OK to
+    use *enable* when not referring to a person.
+    For turning on or activating an option or feature, use *enable* or
+    *[turn on](#word-list--turn-on)* consistently:
+
+    - Use the same term in introductory text as described in the
+      procedure.
+    - Use the same term throughout the document unless there's a
+      difference in the UI elements for different procedures.
+    Recommended: To enable the API,
+    click the toggle.
+    Recommended: Enable the API for your
+    project.
+    For making it feasible to do something, use *lets you*.
+    Recommended: The API lets you detect
+    features in images.
+    Not recommended: The API enables you
+    to detect features in images.
+    Not recommended: The API allows you
+    to detect features in images.
+    In Google Workspace documentation, if possible, use
+    *turn on* or *on* instead. If referring to the state of a UI element, use
+    *available*.
+<a id="word-list--endpoint"></a>
+
+**endpoint**
+    Not *end point*.
+<a id="word-list--enter"></a>
+
+**enter**
+    Use *enter* to refer to the user entering text. If it's important to
+    not press `Enter`, explicitly say so. See also
+    [*type*](#word-list--type).
+    Recommended: In the **Owner** box,
+    enter your name.
+    Recommended: In the **Size** box,
+    type a font size.
+<a id="word-list--ephemeral-external-ip-address"></a>
+
+**ephemeral external IP address**
+    Don't use *ephemeral IP address* or *external IP address* to
+    refer to ephemeral external IP addresses.
+<a id="word-list--error-prone"></a>
+
+**error-prone (adjective)**
+    Hyphenate. Lowercase except at the beginning of a sentence,
+    heading, or list item.
+<a id="word-list--etc"></a>
+
+**Avoid: etc.**
+    Avoid using *etc.*, *and so forth*, and *and so on*
+    wherever possible. If you really need to use one, use *etc.*
+    Always include the period, even if a comma follows immediately after.
+    Recommended: Your app might experience
+    problems such as instability or high latency.
+    Recommended: Your app might experience
+    problems, including instability or high latency.
+    Not recommended: Your app might
+    experience instability, high latency, and so on.
+    Not recommended: Your app might
+    experience instability, high latency, etc.
+    Not recommended: If your app
+    experiences instability, high latency, etc., follow these steps:
+<a id="word-list--eventually"></a>
+
+**Avoid: eventually**
+    Avoid in timeless documentation because this word can become outdated. The
+    word can also prematurely disclose product or feature strategy or
+    inappropriately imply that a product or feature might change.
+    See also
+    [future](#word-list--future) and [soon](#word-list--soon).
+    Recommended: This version of the SDK
+    is deprecated.
+    Not recommended: This version of the
+    SDK is deprecated and eventually will be no longer supported.
+    For more information, see
+    [Timeless documentation](#timeless-documentation).
+<a id="word-list--execute"></a>
+
+**execute**
+    Verb commonly used to refer to function calls, SQL queries, and other processes. When the meaning
+    is the same, use the simpler word *run* instead. If you need to use a more precise term
+    for your context, use that term.
+<a id="word-list--expander-arrow"></a>
+
+**expander arrow**
+    The UI element used to expand or collapse a section of navigation or
+    content. If you describe this element, use the terms *expander arrow*
+    and *expandable section*
+    Don't use terms like *expando* or *zippy*.
+<a id="word-list--exploit"></a>
+
+**exploit**
+    Don't use *exploit* to mean "use."
+    Only use *exploit* in the negative sense, such as to describe
+    *exploiting a security vulnerability*.
+<a id="word-list--external-vpn-gateway"></a>
+
+**external VPN gateway**
+    Write *external* and *gateway* all lowercase except at the
+    beginning of a sentence, heading or list item.
+<a id="word-list--extract"></a>
+
+**extract**
+    Use instead of *unarchive*, *uncompress*, *untar*, or *unzip*.
+
+<a id="word-list--letter-f"></a>
+
+##### F
+
+<a id="word-list--failover"></a>
+
+**fail over (verb), failover (noun, adjective)**
+<a id="word-list--fat"></a>
+
+**Don't use: fat**
+    Don't use. Instead, use a precise modifier that conveys the appropriate
+    meaning. For example, use *high-capacity network connection* instead
+    of *fat connection* or *full-featured client* instead of *fat
+    client*.
+    Instead of using fat in a negative sense, such as *trim the fat*,
+    refer in a more concrete manner to the *removal of unused items*.
+    OK to use as an acronym when referring to file allocation table (FAT).
+<a id="word-list--female-adapter"></a>
+
+**Don't use: female adapter**
+    Don't use. Instead, use a genderless word like *socket*.
+<a id="word-list--fhir"></a>
+
+**Fast Healthcare Interoperability Resources (FHIR)**
+    Refer to *a FHIR* (pronounced "a fire," as in "a FHIR store"), not *an FHIR*.
+    For more information, see
+    [Indefinite articles before abbreviations](#abbreviations--articles).
+<a id="word-list--filename"></a>
+
+**filename**
+    Not *file name*
+<a id="word-list--file-system"></a>
+
+**file system**
+    Not *filesystem*.
+<a id="word-list--fill-in"></a>
+
+**fill in; fill out**
+    Use *fill in* when referring to entering information in individual
+    fields.
+    Use *fill out* when referring to completing an entire form.
+        Recommended: Fill out the
+            questionnaire. Be sure to fill in the required fields.
+<a id="word-list--final-solution"></a>
+
+**Don't use: final solution**
+    Don't use. Instead, use *solution* as a standalone term or, depending
+    on the context, *definitive*, *optimal*, *best*, or *last
+    solution*.
+<a id="word-list--fintech"></a>
+
+**fintech**
+    Write out on first mention: *financial technology (fintech)*. Don't
+    use *FinTech* or *fin-tech*.
+<a id="word-list--firewalls"></a>
+
+**firewalls**
+    Don't use in Compute Engine or networking documentation. Instead, use
+    *firewall rules*.
+    Exception: If you're explaining how firewall rules work, you can explain
+    that every network has an implied virtual distributed firewall.
+    Outside of Compute Engine or networking documentation, the term
+    *firewalls* is acceptable.
+<a id="word-list--first-class"></a>
+
+**Don't use: first class, first-class, first-class citizen**
+    Don't use *first class* or *first-class citizen*. Instead, use
+    another term that's appropriate for the context, such as *higher-order*,
+    *anonymous*, or *nested*, or loosely describe the specific
+    characteristics or features of the entity, resource, language, or framework.
+    Recommended: These widgets have full access to the event system and lifecycle hooks.
+    Not recommended: The widgets are first-class components in the UI framework.
+    Recommended: Virtual machines are higher-order resources that can participate in resource groups and are integrated in a variety of identity, networking, and storage services.
+    Not recommended: Virtual machines are treated as first-class resources across the identity, networking, and storage services.
+    For more information, see
+    [Write inclusive documentation](#inclusive-documentation).
+<a id="word-list--following"></a>
+
+**following**
+    It's not necessary to use a noun after *following* unless it helps
+    provide clarity and enables accessibility. See [Tables](#tables--table-placement).
+    Recommended: ... in the following
+    code sample ...
+    Recommended: ... in the following
+    table ...
+    Recommended: ... do the following:
+    ...
+<a id="word-list--foo"></a>
+
+**Avoid: foo**
+    Avoid when possible even though it's a common term in the developer
+    community. Instead, use a clearer and more meaningful placeholder name.
+<a id="word-list--for-example"></a>
+
+**for example**
+    When you introduce an example using the phrase *for example*, follow the phrase by a comma. For clarity, when
+    introducing an example, separate the example using dashes, commas, or parentheses from the
+    rest of the sentence as appropriate, or introduce the example in a separate sentence.
+    Recommended: Enter a name for the instance—for example, `my-instance-99`.
+    Recommended: Enter a six-digit hex number (for example, `228B22`), and then click **OK**.
+    Recommended: Enter a six-digit hex number, and then click **OK**. For example, if you want the color forest
+    green, enter `228B22`.
+    For more information, see [Format examples](#format-examples).
+<a id="word-list--for-instance"></a>
+
+**Avoid: for instance**
+    Don't use the phrase *for instance* to introduce examples to avoid confusion with the
+    noun *instance*. Instead, use *for example*, *like*, or *such as*. For
+    more information, see [for example](#word-list--for-example).
+<a id="word-list--frontend"></a>
+
+**frontend**
+    Not *front-end* or *front end*.
+<a id="word-list--functionality"></a>
+
+**Avoid: functionality**
+    Use with caution. With respect to hardware or software,
+    *functionality* refers to a set of associated functions or
+    capabilities and how they work. However, the word is sometimes overused,
+    especially when the intended meaning is *capabilities* or
+    *features*.
+<a id="word-list--future"></a>
+
+**Avoid: future, in the future**
+    Avoid in timeless documentation because this word or phrase can become
+    outdated.
+    See also [eventually](#word-list--eventually) and [soon](#word-list--soon). For more
+    information, see [Timeless
+    documentation](#timeless-documentation).
+
+<a id="word-list--letter-g"></a>
+
+##### G
+
+<a id="word-list--gigabytes-per-second"></a>
+
+**GBps**
+    Short for *gigabytes per second*. By convention, we don't use
+    *GB/s*. For more information, see [Units of measurement](#units-of-measure).
+<a id="word-list--gbps"></a>
+
+**Gbps**
+    Short for *gigabits per second*. By convention, we don't use
+    *Gb/s*. For more information, see [Units of measurement](#units-of-measure).
+<a id="word-list--gcloud"></a>
+
+**(Cloud) `gcloud` CLI**
+    Use the full name *Google Cloud CLI* the first time that you mention
+    the product on a page.
+<a id="word-list--gender"></a>
+
+**Don't use: gender-neutral he, him, or his (or she or her)**
+    Don't use. Instead, use the singular *they* (see [Jane Austen and other famous authors violate what everyone learned in
+    their English class](http://www.pemberley.com/janeinfo/austheir.html)). Don't use *he/she* or *(s)he* or other
+    such punctuational approaches. For more information, see
+    [Pronouns](#pronouns).
+<a id="word-list--generative-ai"></a>
+
+**generative AI**
+    Spell out *generative*. Use sentence case.
+    Don't use *gen AI* or *Gen AI*.
+    Don't hyphenate *generative AI* as an adjective unless you must do
+    so for clarity. See also [AI](#word-list--ai).
+<a id="word-list--ghetto"></a>
+
+**Don't use: ghetto**
+    Don't use. Instead use more precise terms like *clumsy*,
+    *workaround*, or *inelegant* to refer to code that isn't in a
+    production-ready state.
+<a id="word-list--gimp"></a>
+
+**Don't use: gimp, gimpy**
+    Don't use. Instead, use precise, non-figurative language to refer to a
+    deficiency in a component.
+    OK to use in reference to companies, tools, software packages, and other
+    entities that use the term in their names.
+<a id="word-list--gke-node"></a>
+
+**GKE node**
+    Use when first introducing GKE nodes on a given page. For subsequent
+    mentions, you can use *node*. A GKE node is a worker machine that
+    runs containerized applications and other workloads. The machine is a
+    Compute Engine VM that GKE creates during cluster creation. See also [virtual machine (VM) instance](#word-list--virtual-machine-instance).
+<a id="word-list--google"></a>
+
+**Avoid: Google, Googling**
+    Don't use as a verb or gerund. Instead, use *search with Google*.
+<a id="word-list--google-account"></a>
+
+**Google Account, Google Accounts**
+    Capitalize *Account*.
+<a id="word-list--google-api-client-library"></a>
+
+**Google API Client Library for LANGUAGE (Java, .NET, etc.)**
+    On second and subsequent use, you can abbreviate to
+    *LANGUAGE client library*.
+<a id="word-list--google-api-console"></a>
+
+**Don't use: Google API Console, Google APIs Console**
+    Don't use. For more information, see [console](#word-list--console).
+<a id="word-list--gcp"></a>
+
+**Google Cloud**
+    Not *GCP*, *Cloud Platform*, or *Cloud*.
+<a id="word-list--google-cloud-platform-console"></a>
+
+**Google Cloud console**
+    If you're only discussing the Google Cloud console, it's OK to shorten to
+    *the console* after first use on a given page.
+    Use *the* before the console name. For more information, see [console](#word-list--console).
+<a id="word-list--gcp-project-id"></a>
+
+**Google Cloud project ID**
+    Not *Cloud project ID* or *GCP project ID*. You can also
+    shorten to *project ID*, but be aware that that term is ambiguous in
+    some contexts.
+<a id="word-list--google-developers-console"></a>
+
+**Don't use: Google Developers Console**
+    Don't use. For more information, see [console](#word-list--console).
+<a id="word-list--google-io"></a>
+
+**Google I/O**
+    Not *I-O* or *IO*.
+<a id="word-list--google-play-services"></a>
+
+**Google Play services**
+    Write *services* in lowercase.
+<a id="word-list--google-play-services-SDK"></a>
+
+**Google Play services SDK**
+    Write *services* in lowercase.
+<a id="word-list--grandfather-clause"></a>
+
+**Don't use: grandfather clause, grand-father clause, grand father clause**
+    Don't use. See [grandfathered](#word-list--grandfathered).
+<a id="word-list--grandfathered"></a>
+
+**Don't use: grandfathered**
+    Don't use to refer to something that is allowed to violate a rule because
+    it predates the rule. Instead, use an adjective like *legacy* or
+    *exempt* or a verb like *made an exception*.
+    Recommended: The app is exempt because
+    it was released before the new requirements were announced.
+    Not recommended: The app is
+    grandfathered in because it was released before the new requirements were
+    announced.
+<a id="word-list--gray-box"></a>
+
+**Avoid: gray-box, grey-box**
+    Avoid using *gray-box*, *graybox*, or *gray box* to
+    describe testing.
+    To refer to testing that's a combination of clear and opaque testing
+    methods, describe exactly what it's doing.
+    If you need to refer to this type of testing after you describe it,
+    consider using a more precise term for clarity, such as *translucent-box
+    testing*.
+<a id="word-list--grayed-out"></a>
+
+**Don't use: grayed-out, greyed-out, gray out, grey out**
+    Don't use. Instead, use *unavailable*.
+<a id="word-list--grayhat"></a>
+
+**Don't use: grayhat, greyhat, gray hat, grey hat**
+    Don't use. Follow the guidance for [black hat](#word-list--blackhat) when
+    referring to someone violating rules or laws.
+<a id="word-list--graylist"></a>
+
+**Don't use: graylist, greylist, gray list, grey list, gray-list, grey-list**
+    Don't use. See [blacklist](#word-list--blacklist).
+<a id="word-list--graylisted"></a>
+
+**Don't use: graylisted, greylisted, gray listed, grey listed, gray-listed, grey-listed**
+    Don't use. See [blacklist](#word-list--blacklist).
+<a id="word-list--graylisting"></a>
+
+**Don't use: graylisting, greylisting, gray listing, grey listing, gray-listing, grey-listing**
+    Don't use. See [blacklist](#word-list--blacklist).
+<a id="word-list--gsutil"></a>
+
+**(Cloud) `gsutil`**
+    In the Google Cloud context, use code font for both the name of the
+    command-line utility and the command.
+<a id="word-list--guru"></a>
+
+**Don't use: guru**
+    If possible, use a more precise term. For example, if you mean
+    *expert* or *teacher*, use those terms.
+<a id="word-list--guys"></a>
+
+**Don't use: guys, you guys**
+    When referring to a group of people use non-gendered language, such as
+    *everyone* or *folks*.
+<a id="word-list--gypsy"></a>
+
+**Don't use: gypsy**
+    Don't use. To refer to the people, use *Romani*, *Roma*, or
+    *Traveller*, as appropriate for the specific group you're referring
+    to. In place of metaphorical uses of the term, use more precise phrases.
+
+<a id="word-list--letter-h"></a>
+
+##### H
+
+<a id="word-list--hamburger"></a>
+
+**Don't use: hamburger, hamburger menu**
+    Don't use. Instead use the `aria-label` for that particular
+    icon. For example, **Menu**.
+    For more information, see
+    [Buttons and icons](#ui-elements--buttons).
+<a id="word-list--hands-off"></a>
+
+**Don't use: hands off, hands-off**
+    Use a less figurative phrase, such as *automated*. If you're
+    referring to a group that doesn't do anything during a process, write a
+    description.
+<a id="word-list--hands-on"></a>
+
+**Don't use: hands on, hands-on**
+    Use a less figurative phrase, such as *customizable*, or write a
+    description of the activity.
+<a id="word-list--hang"></a>
+
+**Don't use: hang, hung**
+    Don't use to refer to a computer or system that is not responding.
+    Instead, use *stop responding* or *not responding*. For more
+    information, see [Avoid figurative
+    language](#inclusive-documentation--figurative-language).
+<a id="word-list--happiness"></a>
+
+**happiness and satisfaction**
+    Use *happiness* when referring to a customer's perception of a
+    site's reliability. Use *satisfaction* when referring to whether the
+    site meets the customer's needs.
+    Site reliability engineering (SRE) content generally refers to
+    measuring *customer happiness* instead of *customer
+    satisfaction*. The two phrases are not equivalent.
+    The distinction the SRE documentation makes is between satisfying a need
+    (a dispassionate act) and establishing an emotional response (creating
+    happiness). Although it is difficult to measure happiness precisely, SRE
+    uses [service level indicators
+    (SLIs)](#word-list--service-level-indicator) to quantify user perception. For example, a customer might feel
+    a "need" to watch a show on TV. If the show is available, the customer's
+    need is satisfied. But if playback is slow or choppy, the customer might
+    not be happy.
+    For more information about SRE and measuring reliability, see [The Happiness Test](https://www.coursera.org/lecture/site-reliability-engineering-slos/the-happiness-test-ELmSr).
+<a id="word-list--hardcode"></a>
+
+**hardcode (verb), hardcoded (adjective)**
+    Don't hyphenate.
+<a id="word-list--he"></a>
+
+**Avoid: he, him, his**
+    Don't use a gendered pronoun except for a specific individual of known
+    gender. Use *they* and *their* for the general singular pronoun.
+<a id="word-list--healthcare"></a>
+
+**healthcare**
+    Not *health care* or *health-care*.
+<a id="word-list--health-check"></a>
+
+**Avoid: health check**
+    Use with caution. When describing an action taken for a computer system,
+    only use the term *health check* if this is the term that appears in
+    the interface. Be certain to remove any ambiguity regarding whether the
+    term refers to health in the medical sense.
+    Use detailed, non-figurative language as much as possible, such as
+    referring to a node *being responsive* instead of referring to a node
+    being healthy.
+<a id="word-list--healthy"></a>
+
+**Avoid: healthy**
+    Don't use. See [health check](#word-list--health-check).
+<a id="word-list--high-availability"></a>
+
+**high availability (noun), high-availability (adjective)**
+    Spell as *high availability* when used as a noun and as *high-availability* when
+    used as an adjective. See also [load balancing (noun), load-balancing (adjective)](#word-list--load-balancing).
+    Lowercase except when part of a product name, but OK to abbreviate as
+    *HA* after first use.
+<a id="word-list--higher"></a>
+
+**higher**
+    Don't use for a range of version numbers. Instead, use [*later*](#word-list--later).
+    Don't use to refer to a position in a document. Use *earlier* or
+    *preceding*.
+    Don't use to refer to a position in the UI. Instead, write instructions
+    that avoid directional language. For more information, see [Writing accessible documentation](#accessibility).
+    In Android documentation, use
+    *higher* for a range of version numbers, not *later*.
+    A release with the highest version number might not be the latest version.
+    For example, if version 2.0 of an operating system receives a bug-fix
+    update after version 3.0 has been released, then version 2.0.1 might be
+    the latest version, even though its version number is lower than 3.0.
+<a id="word-list--high-performance-computing"></a>
+
+**high performance computing (HPC)**
+    Don't hyphenate. Lowercase except at the beginning of a sentence,
+    heading, or list item.
+<a id="word-list--hit"></a>
+
+**Don't use: hit**
+    Don't use as a synonym for *click*, *press*, or *type*.
+<a id="word-list--hold-the-pointer-over"></a>
+
+**hold the pointer over**
+    Only use this verb phrase in the following cases:
+
+    - When the user needs to hold their mouse over a UI element, but not
+      click the UI element. This action involves waiting for the UI to
+      react—for example, waiting for a tooltip to open or waiting for a
+      submenu to open.
+    - When the duration of time is important.
+
+    The phrase *point to* is more common.
+    See also [point to](#word-list--point-to).
+    Recommended: In the **Admin**
+    menu, hold the pointer over **File**, and then click **New**.
+    Not recommended: In the **Admin**
+    menu, hover over **File**, and then click **New**.
+<a id="word-list--holiday"></a>
+
+**Avoid: holiday, the holidays**
+    Don't use to refer to the end of the year. Instead, refer to specific
+    quarters or months.
+<a id="word-list--home-screen"></a>
+
+**(Android) home screen**
+    Two words in Android contexts; not *homescreen* or
+    *home-screen*.
+<a id="word-list--hostname"></a>
+
+**hostname**
+    Not *host name*.
+<a id="word-list--hot"></a>
+
+**hot**
+    When possible, avoid [jargon](#jargon) like *hot failover*,
+    *hot standby*, and *hot spare*. If you use one of these phrases,
+    define it on first use and use it consistently throughout the document. However, see
+    [hotspot](#word-list--hotspot).
+<a id="word-list--hotspot"></a>
+
+**hotspot**
+    In databases, *hotspots* occur when a small number of nearby rows are
+    accessed frequently in a short period of time, causing CPU spikes and
+    affecting performance. Use *hotspot* and *hotspots* as nouns.
+    Don't use verb and gerund forms such as *hotspotting*, because they
+    translate less consistently.
+    When you use *hotspot*, define it the first time that you use it on
+    a page as you normally do with jargon.
+    Recommended: Hotspots in one table
+    can affect the performance of other tables.
+    Not recommended: Hotspotting in one
+    table can affect the performance of other tables.
+<a id="word-list--housekeeping"></a>
+
+**Don't use: housekeeping, house keeping, house-keeping**
+    Don't use. Instead, use less figurative and more precise terms, such as
+    *maintenance* and *cleanup*.
+<a id="word-list--hover"></a>
+
+**Don't use: hover**
+    Don't use. Instead use [*hold the
+    pointer over*](#word-list--hold-the-pointer-over).
+<a id="word-list--https"></a>
+
+**HTTPS**
+    Not *HTTPs*.
+
+<a id="word-list--letter-i"></a>
+
+##### I
+
+<a id="word-list--iaas"></a>
+
+**IaaS**
+    Write out on first mention: *infrastructure as a service (IaaS)*.
+<a id="word-list--iam"></a>
+
+**IAM**
+    When referring to the Google Cloud product, spell it out on first use:
+    *Identity and Access Management (IAM)*.
+    When referring to UI text, write this term the way it's written in the UI.
+    When referring to the general practice of identity and access management,
+    spell it out in lowercase on first use and include a parenthetical
+    comment:
+    Recommended: Identity and access
+    management (generally referred to as *IAM*) is the practice of
+    granting the right individuals access to the right resources for the
+    right reasons.
+<a id="word-list--id"></a>
+
+**ID**
+    Not *Id* or *id,* except in string literals or enums.
+    In some contexts, it's best to spell out as *identifier* or
+    *identification*.
+<a id="word-list--ie"></a>
+
+**Don't use: i.e.**
+    Don't use. Instead, use phrases like *that is*. Many people confuse
+    *e.g.* and *i.e.*
+<a id="word-list--if"></a>
+
+**if**
+    Wondering whether to use *if* or *whether*? See [whether](#word-list--whether).
+    Although it is common in casual usage to omit the word *then* in *if...then*
+    statements, you should include helper words like *then* in technical documentation. For
+    more information, see
+    [Use clear, precise, and unambiguous language](https://developers.google.com/style/translation#clear-language).
+<a id="word-list--image"></a>
+
+**image**
+    *Image* by itself doesn't localize well because of its many meanings. Consider adding
+    context—for example, *disk image* or *container image*.
+<a id="word-list--impact"></a>
+
+**Avoid: impact**
+    Use only as a noun. Instead of writing that something *has an
+    impact*, use the word *affect*.
+    Recommended: This issue affects
+    user experience.
+    Acceptable: This issue has an impact
+    on user experience.
+    Not recommended: This issue impacts
+    user experience.
+<a id="word-list--index"></a>
+
+**index**
+    Use the plural *indexes* unless there is a domain-specific reason
+    (for example, a mathematical or financial context) to use *indices*.
+<a id="word-list--ingest"></a>
+
+**ingest**
+    Use *import*, *load*, or *copy* when referring to simple movement of data. Use
+    *ingest* only when referring to such operations that also involve significant processing
+    of the data.
+<a id="word-list--ingress"></a>
+
+**ingress**
+    When referring to the networking term, use lowercase. When referring
+    to the GKE term or API, capitalize *Ingress*.
+<a id="word-list--in-order-to"></a>
+
+**Avoid: in order to**
+    Avoid *in order to*; instead, use *to*.
+    Use *in order to* when needed to clarify meaning or to make
+    something easier to read.
+    Recommended: You can use
+    monitoring to help identify issues.
+    Not recommended: You can use
+    monitoring in order to help identify issues.
+    Recommended: The infrastructure is
+    required in order to support search.
+    Not recommended: The infrastructure
+    is required to support search.
+<a id="word-list--inline"></a>
+
+**inline**
+    One word as an adjective, *inline*, not *in line* or
+    *in-line*.
+<a id="word-list--instance-group"></a>
+
+**instance group**
+    Don't abbreviate to *IG*. See also [managed instance
+    group](#word-list--mig).
+<a id="word-list--intercluster"></a>
+
+**intercluster**
+    Use unhyphenated *intercluster*, not *inter-cluster*.
+<a id="word-list--interconnect-attachment"></a>
+
+**interconnectAttachment**
+    Use when referring to the API. Otherwise, use [*VLAN attachment*](#word-list--vlan).
+<a id="word-list--interconnect-connection"></a>
+
+**Interconnect connection**
+    Only use *Interconnect connection* relative to a product as follows:
+
+    - CDN Interconnect connection
+    - Cloud Interconnect connection
+    - Dedicated Interconnect connection
+    - Partner Interconnect connection
+
+    OK to use *connection* on subsequent mentions.
+    When you're referring to a Google Cloud product, always specify the
+    product name. Don't use *Interconnect* or *interconnect* as
+    standalone terms, and don't use generic terms like *cloud interconnect
+    connection* or *cross-connect*.
+<a id="word-list--interconnect-connection-location"></a>
+
+**Interconnect connection location**
+    Only refer to an *Interconnect connection location* in context of a
+    specific product, for example *CDN Interconnect*.
+    OK to also use *colocation facility*.
+<a id="word-list--interconnect-type"></a>
+
+**Don't use: interconnect type**
+    Don't use. Instead, use *connection type*. Examples of connection
+    types are a *dedicated connection* or a *connection provided by a
+    service provider*.
+<a id="word-list--interface"></a>
+
+**interface**
+    OK to use as a noun.
+    Don't use as a verb. Instead, use *interact*, *talk*,
+    *speak*, *communicate*, or other similar terms.
+<a id="word-list--internal-dns"></a>
+
+**internal DNS**
+    Write *internal* all lowercase except at the beginning of a
+    sentence, heading, or list item.
+<a id="word-list--idn"></a>
+
+**Internationalized Domain Name (IDN)**
+    Write out and capitalize each word on first use. OK to abbreviate as
+    *IDN* after first use.
+<a id="word-list--internet"></a>
+
+**internet**
+    Lowercase except at the beginning of a sentence,
+    heading, or list item.
+<a id="word-list--ike"></a>
+
+**Internet Key Exchange (IKE)**
+    Write out and capitalize each word on first use. OK to abbreviate
+    *IKE* after first use.
+<a id="word-list--io"></a>
+
+**I/O (see also [Google I/O](#word-list--google-io))**
+    Not *I-O* or *IO*.
+<a id="word-list--iot"></a>
+
+**IoT**
+    OK to use as an abbreviation for *Internet of Things*. Note
+    the lowercase *o*.
+<a id="word-list--ipsec"></a>
+
+**IPsec**
+    Not *IPSec* or *IPSECShort*.
+    Short for *Internet Protocol Security*. No need to spell out on
+    first mention.
+
+<a id="word-list--letter-j"></a>
+
+##### J
+
+<a id="word-list--jank"></a>
+
+**Avoid: jank, janky**
+    Use only to refer to a glitch or problem with graphics that is caused by a loss of data or
+    inadequate refresh rate. Don't use otherwise. Use a less figurative term to refer to something
+    of poor or unreliable quality.
+<a id="word-list--just"></a>
+
+**Avoid: just**
+    Avoid. Usually, *just* is a filler word that you can delete without
+    affecting your meaning.
+    Recommended: BigQuery skips the row.
+    Not recommended: BigQuery just skips
+    the row.
+    If your meaning is unclear without *just*, then use a more specific
+    term such as *only*, *instead*, or *previously*, or revise
+    your language to be more specific. (Even if one of these replacement
+    terms fits, you often don't need it.)
+    Recommended: You can run DML
+    statements in the same way that you'd run a `SELECT`
+    statement.
+    Not recommended: You can run DML
+    statements just as you'd run a `SELECT` statement.
+    Recommended: Let a user query only
+    the table without full dataset access.
+    Recommended: Let a user query the
+    table without full dataset access.
+    Not recommended: Let a user query
+    just the table without full dataset access.
+    Sometimes, *just* is useful for conveying that one approach is
+    simpler than another. In those cases, use *just* instead of
+    [*simply*](#word-list--simple).
+    Recommended: Use the namespace ID
+    `namespace:example-kind` or just `example-kind`.
+
+<a id="word-list--letter-k"></a>
+
+##### K
+
+<a id="word-list--k8s"></a>
+
+**Don't use: k8s**
+    Don't use. Instead, use *Kubernetes*.
+<a id="word-list--kilobytes-per-second"></a>
+
+**KBps**
+    Short for *kilobytes per second*. By convention, we don't use
+    *KB/s*. For more information, see [Units of measurement](#units-of-measure).
+<a id="word-list--kbps"></a>
+
+**Kbps**
+    Short for *kilobits per second*. By convention, we don't use
+    *Kb/s*. For more information, see [Units of measurement](#units-of-measure).
+<a id="word-list--kebab"></a>
+
+**Don't use: kebab, kabob, kebab menu, kabob menu**
+    Don't use. Instead use the `aria-label` for that particular
+    icon. For example,
+    **More**. For more information, see
+    [Buttons and icons](#ui-elements--buttons).
+<a id="word-list--kebab-case"></a>
+
+**Don't use: kebab case, kabob case, kebab-case, kabob-case**
+    Don't use. Instead, use *dash-case*.
+<a id="word-list--key"></a>
+
+**key**
+    Don't use as an adjective in the sense of *crucial* or
+    *important*.
+    If you use *key* as a noun, specify which kind of key you're
+    referring to on first mention, because there are many kinds of
+    keys in technical contexts.
+<a id="word-list--key-pair"></a>
+
+**key pair**
+    A pair of keys, such as a public key and a private key. Contrast with
+    *key-value pair*, which refers to a pairing that specifies a value
+    for a variable (as in configuration files).
+<a id="word-list--key-ring"></a>
+
+**key ring**
+    Use instead of *keyring* (without the space) when referring to a
+    grouping of Cloud KMS keys.
+<a id="word-list--key-value"></a>
+
+**key-value pair**
+    Use instead of *key/value pair* or *key value pair*.
+<a id="word-list--kill"></a>
+
+**Avoid: kill**
+    Avoid when possible. Instead, use words like *stop*, *exit*,
+    *cancel*, or *end*. For exceptions to this rule, see
+    [Documenting command-line
+    syntax](#code-syntax--linux-signals).
+
+<a id="word-list--letter-l"></a>
+
+##### L
+
+<a id="word-list--lame"></a>
+
+**Don't use: lame**
+    Don't use. Instead, use precise, non-figurative language to refer to a
+    deficiency in a component.
+<a id="word-list--later"></a>
+
+**later**
+    Use for a range of version numbers, not *higher*.
+    Recommended: Use version 2.2 or
+    later.
+    Not recommended: Use version 2.2 or
+    higher.
+    Not recommended: Use version 2.2+.
+    A release with the highest version number might not be the latest version.
+    For example, if version 2.0 of an operating system receives a bug-fix
+    update after version 3.0 has been released, then version 2.0.1 might be
+    the latest version, even though its version number is lower than 3.0.
+    In Android documentation, don't use
+    *later* for a range of version numbers. Instead, use *higher*.
+    When referring to a position in a document, use *later* or
+    *following*, not *below*.
+<a id="word-list--latest"></a>
+
+**Avoid: latest**
+    Avoid in timeless documentation because this word can become outdated.
+    If you must use *latest*, give the reader a reference
+    point—for example, a version number or release date.
+    Recommended: To help keep your
+    system secure, install the latest version of the tools.
+    Recommended: The June 2021 release
+    includes the latest tools that help secure your system.
+    Not recommended: The product includes
+    the latest tools that help secure your system.
+    For more information, see
+    [Timeless documentation](#timeless-documentation).
+<a id="word-list--learnings"></a>
+
+**Don't use: learnings**
+    Don't use. Instead, refer to *knowledge* or *things that you
+    learned*.
+<a id="word-list--left-nav"></a>
+
+**Don't use: left-nav, right-nav**
+    Don't use directional language. For more information, see
+    [Writing accessible documentation](#accessibility).
+    If referring to applications, use *[navigation menu](#ui-elements--term-navigation-menu)*.
+    If referring to navigational elements for documentation, use *content
+    navigation menu*.
+<a id="word-list--legacy"></a>
+
+**legacy**
+    If possible, use a more precise term. If you do use *legacy*,
+    include or point to a definition to clarify what you mean in the current
+    context. Don't use *legacy* with any sort of pejorative
+    connotation.
+<a id="word-list--lets"></a>
+
+**Don't use: let's (as a contraction of *let us*)**
+    Don't use if at all possible.
+    Not recommended: Let's click the
+    **OK** button now.
+<a id="word-list--loa-cfa"></a>
+
+**Letter of Authorization and Connecting Facility Assignment (LOA-CFA)**
+    Write out and capitalize each word on first use. OK to abbreviate as
+    *LOA-CFA* after first use.
+<a id="word-list--leverage"></a>
+
+**Avoid: leverage**
+    Avoid using if you mean *use*. If possible, use a more precise term.
+    For example, *use*, *build on*, or *take advantage of*.
+<a id="word-list--lifecycle"></a>
+
+**lifecycle**
+    Not *life cycle* or *life-cycle*.
+<a id="word-list--lift-and-shift"></a>
+
+**lift and shift**
+    See [rehost](#word-list--rehost).
+<a id="word-list--like"></a>
+
+**like**
+    It's OK to use *like* for either drawing comparisons (in the sense of *similar to*)
+    or introducing examples (in the sense of *such as*).
+    Recommended: Common I/O operations, like reading files or
+    making network requests, can be asynchronous.
+    Recommended: The new compression algorithm works like a
+    dictionary encoder, replacing repeated strings with shorter codes.
+    See also [such as](#word-list--such-as). For more information, see [Format examples](#format-examples).
+<a id="word-list--limits"></a>
+
+**limits**
+    In an API context, *limit* often refers to usage limits (number of
+    queries allowed per second or per day). Where possible, specify the kind
+    of limit that you mean, such as *usage limit* or *service
+    limit*; the word *limit* can refer to many different kinds of
+    limits, including rules about acceptable use. See also [quota](#word-list--quota).
+<a id="word-list--lint"></a>
+
+**lint**
+    Write both command-line tool name and command in lowercase. Use code font
+    except where inappropriate.
+<a id="word-list--little-endian"></a>
+
+**little-endian**
+    Hyphenate. Lowercase except at the beginning of a sentence,
+    heading, or list item.
+    Recommended: The codebase assumes
+    little-endian byte ordering.
+    Not recommended: The codebase assumes
+    Little Endian byte ordering.
+    Not recommended: The codebase assumes
+    Little-endian byte ordering.
+    Not recommended: The codebase assumes
+    little endian byte ordering.
+<a id="word-list--livestream"></a>
+
+**livestream**
+    Not *live stream*.
+<a id="word-list--load-balancing"></a>
+
+**load balancing (noun), load-balancing (adjective)**
+    Spell as *load balancing* when used as a noun and as *load-balancing* when
+    used as an adjective. See also [high availability (noun),
+    high-availability (adjective)](#word-list--high-availability).
+<a id="word-list--lock-screen"></a>
+
+**(Android) lock screen**
+    Two words in Android contexts; not *lockscreen* or
+    *lock-screen*.
+<a id="word-list--login"></a>
+
+**login (noun or adjective), log in (verb)**
+    For the verb form, *sign in* is generally better.
+    If you're documenting a tool that uses the term *log in*, then use
+    that term.
+<a id="word-list--long-press"></a>
+
+**Don't use: (Android) long press**
+    In Android documentation, don't use. Instead, use *touch & hold*.
+    (Not *touch and hold*.)
+<a id="word-list--lro"></a>
+
+**long-running operation**
+    Not *long running operation*.
+    OK to abbreviate as *LRO* after the first use.
+<a id="word-list--lower"></a>
+
+**lower**
+    Don't use for a range of version numbers. Instead, use [*earlier*](#word-list--earlier).
+    Don't use to refer to a position in a document. Instead, use *later*
+    or *following*.
+    Don't use to refer to a position in the UI. Instead, write instructions
+    that avoid directional language. For more information, see [Writing accessible documentation](#accessibility).
+    In Android documentation, use
+    *lower* for a range of version numbers, not *earlier*.
+
+<a id="word-list--letter-m"></a>
+
+##### M
+
+<a id="word-list--male-adapter"></a>
+
+**Don't use: male adapter**
+    Don't use. Instead, use a genderless word like *plug*.
+<a id="word-list--man-hours"></a>
+
+**Don't use: man hours, manhours, man-hours**
+    Avoid using gendered terms. Instead use terms like *person hours*.
+<a id="word-list--mitm"></a>
+
+**Don't use: man-in-the-middle (MITM)**
+    Avoid using gendered terms. Instead use terms like *on-path
+    attacker* or *person-in-the-middle (PITM)*.
+<a id="word-list--mig"></a>
+
+**managed instance group (MIG)**
+    OK to abbreviate to *MIG* on subsequent mention. See also
+    [instance group](#word-list--instance-group).
+<a id="word-list--manmade"></a>
+
+**Don't use: manmade, man made**
+    Avoid using gendered terms. Instead use a word like *artificial*,
+    *manufactured*, or *synthetic*.
+<a id="word-list--manned"></a>
+
+**Don't use: manned**
+    Avoid using gendered terms. Instead use terms like *staffed* or
+    *crewed*.
+<a id="word-list--manpower"></a>
+
+**Don't use: manpower, man power, man-power**
+    Avoid using gendered terms. Instead use terms like *staff* or
+    *workforce*.
+<a id="word-list--markdown"></a>
+
+**Markdown**
+    Always capitalized, even when you're referring to a nonstandard version.
+<a id="word-list--master"></a>
+
+**Don't use: master**
+    Use with caution. Never use in conjunction with *slave*. Where
+    possible, replace *master* with a specific term that is accurate for
+    the context, such as *primary*, *main*, *original*,
+    *parent*, *initiator*, *driver*, *controller*,
+    *manager*, *mixer*, *aggregator*, *publisher*,
+    *leader*, or *active*.
+
+    Guidance | Recommended | Not recommended || Don't use *master* in conjunction with *slave* in any context. | Cloud SQL primary/replica | Cloud SQL master/slave |
+    | Avoid using *master* where possible. | - GKE control plane - Jenkins controller - root key (in security) - primary key (in databases) | - GKE master plane - Jenkins master - master key (in security) - master key (in databases) |
+    If the command or code that you're documenting uses the literal word
+    *master*, then use this word only in direct reference to the code
+    item ([formatted as code](#code-in-text)), make it clear
+    what you're referring to, and use the new term thereafter.
+    See also [*slave*](#word-list--slave).
+<a id="word-list--material-design"></a>
+
+**Material Design**
+    Capitalize each word in *Material Design*.
+<a id="word-list--matrix"></a>
+
+**matrix**
+    Use the plural *matrixes* unless there is a domain-specific reason
+    (for example, a mathematical context) to use *matrices*.
+<a id="word-list--may"></a>
+
+**may**
+    In general, reserve for official policy or legal considerations.
+    To convey *possibility*, use *can* or *might*
+    instead.
+    To convey *permission*, use *can* instead.
+    See also [can](#word-list--can), [could](#word-list--could),
+    [might](#word-list--might), [must](#word-list--must),
+    [should](#word-list--should), and [would](#word-list--would).
+    For information about clarifying who's performing an action, see
+    [Active voice](#voice).
+<a id="word-list--megabytes-per-second"></a>
+
+**MBps**
+    Short for *megabytes per second*. By convention, we don't use
+    *MB/s*. For more information, see
+    [Units of measurement](#units-of-measure).
+<a id="word-list--mbps"></a>
+
+**Mbps**
+    Short for *megabits per second*. By convention, we don't use
+    *Mb/s*. For more information, see
+    [Units of measurement](#units-of-measure).
+<a id="word-list--media-type"></a>
+
+**media type**
+    In general, use the term [*media type*](https://www.iana.org/assignments/media-types/media-types.xhtml).
+    In contexts where you need to refer to a *content type*—For example, if you mention
+    the `Content-Type` HTTP header—it's okay to use *content type* instead, to avoid
+    confusion. Don't use *MIME type*.
+<a id="word-list--meta"></a>
+
+**meta***
+    See [guidance about hyphens with prefixes](#hyphens--prefixes).
+<a id="word-list--metafeed"></a>
+
+**metafeed**
+    Not *meta-feed*.
+<a id="word-list--metageneration"></a>
+
+**metageneration**
+    Not *meta-generation*.
+<a id="word-list--method"></a>
+
+**method**
+    In programming contexts where *method* refers to a member of a class
+    (as in Java), avoid also using the word generically to mean "approach" or
+    "manner."
+<a id="word-list--metro"></a>
+
+**metropolitan area (metro)**
+    In networking, a *metro* is a city where a colocation facility is
+    located.
+<a id="word-list--microservices"></a>
+
+**microservices**
+    Not *Microservices* or *micro-services*.
+<a id="word-list--might"></a>
+
+**might**
+    Use to convey possibility or an uncertain outcome (for example, "You
+    might be prompted to enter your credentials").
+    See also [can](#word-list--can), [could](#word-list--could),
+    [may](#word-list--may), [must](#word-list--must),
+    [should](#word-list--should), and [would](#word-list--would).
+    For information about clarifying who's performing an action, see
+    [Active voice](#voice).
+<a id="word-list--mime-type"></a>
+
+**Avoid: MIME type**
+    *MIME* stands for "Multipurpose Internet Mail Extensions," and was originally used to
+    refer to email standards.
+    Don't use *MIME* when you mean [*media type*](https://www.iana.org/assignments/media-types/media-types.xhtml).
+    If you feel that might be ambiguous to an audience familiar with the term *MIME*,
+    then you can write *media (MIME) type* for clarity.
+<a id="word-list--mobile"></a>
+
+**Don't use: mobile**
+    Don't use *mobile* as a standalone noun. Instead, specify
+    *mobile phone*, or if you're talking about more than phones, then use
+    *mobile device*.
+<a id="word-list--mobile-data"></a>
+
+**mobile data**
+    Use instead of *cellular data*.
+<a id="word-list--mobile-device"></a>
+
+**mobile device**
+    Use *mobile device* when you're referring to more than phones (for
+    example, tablets and phones). It's OK to use *phone* (without
+    *mobile*) when the context is clear.
+<a id="word-list--mobile-network"></a>
+
+**mobile network**
+    Use instead of *cellular network*.
+<a id="word-list--mobile-phone"></a>
+
+**mobile phone**
+    If you're talking about more than phones, then use *mobile device*.
+    It's OK to use *phone* (without *mobile*) when the context is
+    clear.
+<a id="word-list--mom-test"></a>
+
+**Don't use: mom test**
+    Don't use *mom test*, *grandmother test*, *grandma test*,
+    or *girlfriend test*. Instead, use terms like *beginner user
+    test* or *novice user test*.
+<a id="word-list--monkey"></a>
+
+**Don't use: monkey, monkey test**
+    Don't use *monkey* to refer to people. When referring to tests, refer
+    to the specific function. For example: *automated, random tests*.
+<a id="word-list--multi"></a>
+
+**multi***
+    See [guidance about hyphens with prefixes](#hyphens--prefixes).
+<a id="word-list--multi-cluster"></a>
+
+**multi-cluster**
+    Hyphenate. We generally prefer to close prefixed words, but this is an
+    exception because it's an established term.
+<a id="word-list--multi-region"></a>
+
+**multi-region, multi-regional**
+    Hyphenate when referring to a Google Cloud location that consists of more
+    than one region.
+    You can use *multi-regional* as an adjective in the context of
+    multi-regions, but consider *multi-region* as
+    an attributive noun instead, such as in "The dataset is in the EU
+    multi-region location." Use *multiregional* in other contexts.
+<a id="word-list--multi-service"></a>
+
+**multi-service**
+    Hyphenate. We generally prefer to close prefixed words, but this is
+    an exception because it's an established term.
+<a id="word-list--multi-tenancy"></a>
+
+**multi-tenancy**
+    Hyphenate. We generally prefer to close prefixed words, but this is
+    an exception because it's an established term.
+<a id="word-list--must"></a>
+
+**must**
+    Use to describe a required action or state (for example, "You must have
+    the Editor role"). You can also write *you need* in order to convey a
+    requirement.
+    See also [can](#word-list--can), [could](#word-list--could),
+    [may](#word-list--may), [might](#word-list--might),
+    [should](#word-list--should), and [would](#word-list--would).
+    For information about clarifying who's performing an action, see
+    [Active voice](#voice).
+
+<a id="word-list--letter-n"></a>
+
+##### N
+
+<a id="word-list--na"></a>
+
+**N/A**
+    Not *NA*. Spell out as *not available* or *not applicable*
+    on first reference.
+<a id="word-list--name-server"></a>
+
+**name server**
+    Not *nameserver*.
+<a id="word-list--namespace"></a>
+
+**namespace**
+    Not *name space*.
+<a id="word-list--native"></a>
+
+**Avoid: native**
+    Avoid using *native* to refer to people.
+    When referring to software products, try to use a more precise
+    term—for example, use *built-in* to describe a feature that's
+    part of a product.
+    The term *native* isn't necessarily clear—for example,
+    *cloud-native* could mean that something was written for the cloud,
+    or that it's built in to a cloud platform, or that it currently exists in
+    a cloud platform.
+    Alternatives to a term like *cloud-native* could include:
+    *modern cloud*, *born in the cloud*, *cloud first*, and
+    *cloud-born*.
+<a id="word-list--navigation-bar"></a>
+
+**navigation bar**
+    Don't use to refer to a *navigation menu*. For more information, see
+    [Navigation menu](#ui-elements--term-navigation-menu).
+<a id="word-list--neither"></a>
+
+**neither**
+    Write *neither A nor B*, not *neither A or B*.
+<a id="word-list--network-ip-address"></a>
+
+**Don't use: network IP address**
+    Don't use. Instead, use *internal IP address*.
+<a id="word-list--new"></a>
+
+**Avoid: new, newer**
+    Avoid in timeless documentation because this word can become outdated.
+    *New* also implies that the reader knows the older product and that
+    labeling something as *new* is therefore meaningful.
+    If you must use *new*, give the reader a reference point—for
+    example, a version number or release date.
+    Don't use *newer* to refer to a specific version of a product.
+    Instead, use [*later*](#word-list--later). Make sure that you provide
+    a version number or release date by which to understand *later*.
+
+    In Android documentation, use
+    [*higher*](#word-list--higher) instead of *later*.
+    Recommended: The service's network
+    analysis feature reports on network health.
+    Not recommended: Network analysis, a
+    new feature in the service, reports on network health.
+    For more information, see
+    [Timeless documentation](#timeless-documentation).
+<a id="word-list--ninja"></a>
+
+**Don't use: ninja**
+    Don't use to refer to a person. Instead, use a term such as *expert*.
+    OK to use in reference to companies, tools, software packages, and other
+    entities that use the term in their names.
+<a id="word-list--non"></a>
+
+**non***
+    See [guidance about hyphens with prefixes](#hyphens--prefixes).
+<a id="word-list--nonce"></a>
+
+**Avoid: nonce**
+    Use with caution: this term has a secondary slang meaning that can cause
+    confusion for global readers. Always define the term on first use, and
+    only use it in specific technical contexts such as authentication and
+    blockchain.
+    In end-user documentation and other contexts, use a more descriptive
+    phrase, such as *a number that will be used only once*.
+<a id="word-list--non-key"></a>
+
+**non-key**
+    An exception to our usual preference for closed forms.
+<a id="word-list--noops"></a>
+
+**Don't use: NoOps**
+    Don't use. Instead, use *fully managed*. If you must include the
+    term, define it at first use with language such as *fully managed* or
+    *no operations*, but not *non-operational*. Don't use
+    *noops*.
+    For an instruction that does nothing, use
+    [*no-op*](https://wikipedia.org/wiki/NOP_(code)) or the
+    specific instruction name for your context.
+<a id="word-list--nosql"></a>
+
+**NoSQL**
+    Not *No-SQL* or *No SQL*.
+<a id="word-list--notification-drawer"></a>
+
+**(Android) notification drawer**
+    In Android contexts, don't hyphenate. Lowercase except at the beginning of a sentence,
+    heading, or list item.
+<a id="word-list--now"></a>
+
+**Avoid: now**
+    Avoid when describing features of products or services because this word
+    is implied.
+    If the intent of the text is a comparison between past and present, you
+    can use *now*—for example, "In versions of the tool earlier
+    than 1.10, you could use only the default value, but now you can assign a
+    custom value."
+    Recommended: This feature lets you use
+    combinations of user properties.
+    Not recommended: This feature now lets
+    you use combinations of user properties.
+    For more information, see
+    [Timeless documentation](#timeless-documentation).
+<a id="word-list--nuke"></a>
+
+**Don't use: nuke**
+    Don't use. Instead use *remove* or *attack*. For example, a
+    *denial-of-service attack*.
+
+<a id="word-list--letter-o"></a>
+
+##### O
+
+<a id="word-list--oauth-20"></a>
+
+**OAuth 2.0**
+    Not *OAuth 2*, *OAuth2*, or *Oauth*.
+<a id="word-list--off-the-shelf"></a>
+
+**Don't use: off-the-shelf, commercial off-the-shelf (COTS)**
+    Use more widely understood terms like *ready-made*, *prebuilt*,
+    *standard*, or *default*.
+<a id="word-list--old"></a>
+
+**Avoid: old, older**
+    Don't use to refer to a previous version of a product. Instead, use
+    [*earlier*](#word-list--earlier).
+    Make sure that you provide a version number by which to understand
+    *earlier*.
+    In Android documentation, use
+    [*lower*](#word-list--lower) instead of *earlier*.
+    Recommended: This functionality
+    doesn't work in versions earlier than 1.17.0.
+    Not recommended: This functionality
+    doesn't work in older versions.
+    For more information, see
+    [Timeless documentation](#timeless-documentation).
+<a id="word-list--omnibox"></a>
+
+**Don't use: omnibox**
+    Don't use. Instead, use *address bar*.
+<a id="word-list--once"></a>
+
+**once**
+    If you mean *after*, then use *after* instead of *once*.
+<a id="word-list--on-premises"></a>
+
+**on-premises**
+    Not *on prem*, *on premise*, or *on-premise*. Hyphenate
+    when used as any part of speech.
+    Use to refer to a customer's resources that they manage in their own
+    facilities. Don't use *peer*.
+    It can be acceptable to use *on-premises* as a noun when it would be
+    awkward to repeatedly write out a full phrase like *an on-premises
+    environment*. However, it's preferable to use the more complete phrase
+    whenever possible.
+    Recommended: An on-premises database.
+    Recommended: The database runs
+    on-premises.
+    OK: Moving data from on-premises to
+    Google Cloud.
+<a id="word-list--os"></a>
+
+**OS**
+    OK to use as a shortening of "operating system."
+<a id="word-list--outpost"></a>
+
+**Don't use: outpost**
+    Don't use. Instead, use *channel*.
+    Recommended: social media channels
+<a id="word-list--out"></a>
+
+**Avoid: outside the box, out of the box, out-of-the-box**
+    Avoid using in a figurative way. OK to use literally.
+<a id="word-list--overview-screen"></a>
+
+**Don't use: (Android) overview screen**
+    In Android documentation, don't use. Instead, use *recents screen*.
+
+<a id="word-list--letter-p"></a>
+
+##### P
+
+<a id="word-list--paas"></a>
+
+**PaaS**
+    Write out on first mention: *platform as a service (PaaS)*.
+<a id="word-list--page"></a>
+
+**page**
+    Use *page* to refer to the following:
+
+    - A whole web page, which can include text, images, links, banners, navigational panes,
+      and other features.
+    - A sub-page of a [console](#word-list--console) in particular.
+
+    See also
+    [documentation or document or documents](#word-list--documentation).
+    Recommended: To refresh the page, press `F5`.
+<a id="word-list--parameter"></a>
+
+**parameter**
+    In our API documentation, *parameter* is usually short for *query
+    parameter*; it's a `NAME=VALUE` pair
+    that's appended to a URL in an HTTP `GET` request. In some
+    contexts, however, the term can have other meanings.
+<a id="word-list--parent-child"></a>
+
+**parent-child or parent/child**
+    Not *parent – child* or *parent—child*.
+<a id="word-list--path"></a>
+
+**path**
+    Avoid using *filepath*, *file path*, *pathname*, or *path
+    name* if possible.
+<a id="word-list--peer-gateway"></a>
+
+**peer gateway**
+    Don't use *on-premises gateway* when you mean a *peer gateway*.
+    A peer gateway can be an on-premises device or service or another cloud
+    gateway.
+<a id="word-list--peer-network"></a>
+
+**peer network**
+    Don't use *on-premises network* when you mean a *peer network*.
+    A peer network can be an on-premises network or another cloud network.
+<a id="word-list--peering-zone"></a>
+
+**peering zone**
+    Not *peer zone*.
+<a id="word-list--per"></a>
+
+**per**
+    To express a rate, use *per* instead of the division slash (/),
+    unless space constraints require the use of the slash. For more
+    information, see [Units of
+    measurement](#units-of-measure--rates).
+    Avoid *per* in contexts other than rate units.
+    Recommended: requests per day
+    Recommended: create a policy for each
+    Pod
+    Recommended: according to the style
+    guide
+    Recommended: in response to your
+    request
+    Not recommended: requests/day
+    Not recommended: create a policy per
+    Pod
+    Not recommended: per the style guide
+    Not recommended: as per your request
+<a id="word-list--performant"></a>
+
+**Avoid: performant**
+    Avoid where possible. Instead, use a more precise term.
+    Recommended: an accurate machine
+    learning model
+    Not recommended: a performant machine
+    learning model
+<a id="word-list--persist"></a>
+
+**persist**
+    Don't use as a transitive verb. It's best to avoid using as a verb at all,
+    especially in [passive voice](#voice).
+    Recommended: To make the token
+    persistent ...
+    OK: To make the token persist ...
+    Not recommended: The token is persisted
+    ...
+    Not recommended: To persist the token
+    ...
+<a id="word-list--persistent-disk"></a>
+
+**persistent disk**
+    Not *PD*.
+        Lowercase except at the start of a sentence.
+<a id="word-list--pii"></a>
+
+**personally identifiable information (PII)**
+    Some government agencies use the less common term *personally
+    identifying information*; use this alternate term only in contexts
+    where you're referring to a document that uses this term.
+<a id="word-list--pets-versus-cattle"></a>
+
+**Don't use: pets versus cattle, pets vs. cattle, pets v. cattle**
+    Don't use. Instead, use more precise terms like *persistent versus
+    dynamic* or *manually configured versus automated*. For more
+    information, see
+    [Avoid
+    figurative language](#inclusive-documentation--figurative-language).
+<a id="word-list--plain-text"></a>
+
+**plain text**
+    In most contexts, use *plain text*, but use *plaintext* in a
+    cryptography context.
+<a id="word-list--please"></a>
+
+**Avoid: please**
+    Don't use *please* in the normal course of explaining how to use a
+    product, even if you're explaining a difficult task.
+    Don't use the phrase *please note*.
+    Use *please* only when you're asking for permission or
+    forgiveness—for example, when what you're asking for benefits you,
+    inconveniences a reader, or suggests a potential issue with a product.
+    Recommended: If the issue persists,
+    please contact your account representative.
+    For more information, see
+    [voice and tone](#tone--politeness).
+<a id="word-list--plugin"></a>
+
+**plugin (noun), plug-in (adjective), plug in (verb)**
+    Use the noun form *plugin* when referring to the software component. Use the adjective
+    form *plug-in* when referring to the action of installing a software component. Use the
+    verb form *plug in* when you're describing the process of installing a software
+    component.
+<a id="word-list--pm"></a>
+
+**PM**
+    See [AM, PM](#word-list--am-pm).
+<a id="word-list--point-to"></a>
+
+**point to**
+    Use to refer to the action of pointing the mouse pointer (focus). This
+    action doesn't imply a length of time waiting for the UI to react to user
+    action.
+    This is similar to the action [hold the
+    pointer over (hover)](#word-list--hold-the-pointer-over). In most cases, it's better to use the verb
+    phrase *hold the pointer over* if you want the user to wait for the
+    UI to react.
+<a id="word-list--pojo"></a>
+
+**POJO**
+    If you're not actually writing about a Plain Old Java Object for a Java
+    audience, use *simple object*. You can write *a simple object,
+    similar to a POJO in Java* if that helps your audience.
+<a id="word-list--pop"></a>
+
+**PoP**
+    Acronym for *point of presence*.
+    Recommended: point of presence (PoP)
+    Not recommended: point of presence
+    (POP)
+<a id="word-list--pop-up"></a>
+
+**pop-up, popup**
+    Don't use.
+    To describe a window that appears and asks for, or presents, additional
+    information, use [*dialog*](#word-list--dialog).
+    To describe a menu that rises from an interface (such as a right-click
+    context menu), use *menu*.
+<a id="word-list--populate"></a>
+
+**populate**
+    OK to use if you're writing about a process populating a table or other
+    entity. If you're writing about a person, use *fill in*.
+    Recommended: The SQL command
+    populates the table with sample data.
+    Recommended: When you have finished
+    filling in the form ...
+    Not recommended: When you have
+    finished populating the form ...
+<a id="word-list--port"></a>
+
+**port**
+    Use *listen on* (not *to*).
+<a id="word-list--portal"></a>
+
+**Avoid: portal**
+    Don't use to refer to the Google Cloud console. For more information, see
+    [console](#word-list--console).
+<a id="word-list--possible"></a>
+
+**Don't use: possible**
+    Don't use *possible* or *impossible* to mean *you can* or
+    *you can't*.
+<a id="word-list--postgresql"></a>
+
+**PostgreSQL**
+    If the UI uses the name *Postgres*, it's OK to match the UI. Don't
+    use *PostgreSQL*.
+<a id="word-list--postmortem"></a>
+
+**Avoid: postmortem**
+    Avoid in general usage. Instead, use *retrospective*.
+    In disaster recovery (DR) and DevOps contexts, use *blameless
+    postmortem*.
+<a id="word-list--practitioner"></a>
+
+**practitioner**
+    Avoid using without any supporting information to define the roles that
+    you're referring to.
+    Recommended: The framework describes
+    best practices for architects, developers, administrators, and other cloud
+    practitioners.
+    Not recommended: The framework
+    describes best practices for cloud practitioners.
+<a id="word-list--pre"></a>
+
+**pre***
+    See [guidance about hyphens with prefixes](#hyphens--prefixes).
+<a id="word-list--prebuilt"></a>
+
+**prebuilt**
+    Not *pre-built*.
+<a id="word-list--precapture"></a>
+
+**precapture**
+    Not *pre-capture*.
+<a id="word-list--preemptible"></a>
+
+**preemptible**
+    Not *pre-emptible* or *pre-emptive*.
+<a id="word-list--pre-existing"></a>
+
+**pre-existing**
+    Not *preexisting*.
+<a id="word-list--preferred-pronouns"></a>
+
+**Don't use: preferred pronouns**
+    Don't use. Instead, use *pronouns*.
+<a id="word-list--prerecorded"></a>
+
+**prerecorded**
+    Not *pre-recorded*.
+<a id="word-list--pre-shared-key"></a>
+
+**pre-shared key**
+    Not *preshared key*.
+<a id="word-list--presently"></a>
+
+**Avoid: presently, at present**
+    Avoid because this word or phrase is implied. The word or phrase can also
+    prematurely disclose product or feature strategy or inappropriately imply
+    that a product or feature might change.
+    See also [as of this writing](#word-list--as-of-this-writing) and
+    [currently](#word-list--currently).
+    Recommended: This setting is required.
+    Not recommended: At present, this
+    setting is required.
+    For more information, see
+    [Timeless documentation](#timeless-documentation).
+<a id="word-list--press"></a>
+
+**press**
+    Use when referring to pressing a key or a key combination to cause an
+    action to occur. Also use for mechanical buttons.
+    For on-screen and soft (capacitive) buttons, use *tap*.
+    Recommended: Press
+    `Control+C` (or `Command+C` on macOS).
+<a id="word-list--presubmit"></a>
+
+**presubmit**
+    Not *pre-submit*.
+<a id="word-list--primitive"></a>
+
+**Avoid: primitive**
+    Use with caution. Don't use *primitive* in a disparaging sense.
+<a id="word-list--project"></a>
+
+**(Cloud) project**
+    In Google Cloud documentation, use *Google Cloud project* on first
+    mention and in any context in which there might be ambiguity about what
+    kind of project you're referring to.
+<a id="word-list--pros"></a>
+
+**Don't use: pros**
+    Don't use. Instead, use a more precise term, such as *advantages*.
+
+<a id="word-list--letter-q"></a>
+
+##### Q
+
+<a id="word-list--quick"></a>
+
+**Avoid: quick, quickly**
+    What might be quick for you might not be quick for others. Try
+    eliminating this word from the sentence because usually the same meaning
+    can be conveyed without it.
+<a id="word-list--quota"></a>
+
+**quota**
+    In API contexts, often refers to API usage limits. Where possible, it's
+    best to use a more specific term, such as *usage limit*; the word
+    *quota* means many different things to many different people.
+    In some contexts, such as Google Cloud documentation, the standard term is
+    *quota*, so use that term.
+
+<a id="word-list--letter-r"></a>
+
+##### R
+
+<a id="word-list--rdp"></a>
+
+**RDP**
+    Don't use as a verb. Instead, use *connect using RDP*. If it's
+    clear from context that they're using RDP, it's OK to use *connect*.
+<a id="word-list--re"></a>
+
+**re***
+    See [guidance about hyphens with prefixes](#hyphens--prefixes).
+<a id="word-list--read-only"></a>
+
+**read-only**
+    Not *read only*. Always hyphenate *read-only*.
+<a id="word-list--recents-screen"></a>
+
+**(Android) recents screen**
+    In Android contexts, use instead of *overview screen*.
+<a id="word-list--redline"></a>
+
+**Don't use: redline**
+    Don't use as a verb. Instead, use precise terms appropriate to the
+    context.
+    In the context of editing or providing a review, refer to those actions or
+    to *tracking changes*.
+    In the context of setting priorities and planning work, refer to those
+    actions or to *priority lining*.
+<a id="word-list--regex"></a>
+
+**Don't use: regex**
+    Don't use. Instead, use *regular expression*.
+<a id="word-list--rehost"></a>
+
+**rehost**
+    Use to describe the migration of an app or workload with no changes or
+    minimal changes to that app or workload. Also known as *lift and shift*. For more
+    information, see [Rehost: lift and shift](https://cloud.google.com/architecture/migration-to-gcp-getting-started#rehost_lift_and_shift) in the Cloud Architecture Center.
+    On first mention, associate rehost with lift and shift. Okay to use *rehosting* as needed
+    after first mention.
+    Recommended: You can use this reference architecture to
+    efficiently rehost (lift and shift) on-premises applications to the cloud.
+    Recommended: The first step to modernization is to rehost
+    your application in the cloud (also known as lift and shift).
+    Don't use *the forklift approach*.
+<a id="word-list--repo"></a>
+
+**Don't use: repo**
+    Don't use. Instead, use *repository*.
+<a id="word-list--rest"></a>
+
+**Don't use: Representational State Transfer**
+    Don't use. To people unfamiliar with REST, this acronym expansion is
+    meaningless; it's better to refer to it as REST and not explain what it
+    stands for.
+<a id="word-list--reservation"></a>
+
+**Don't use: reservation, off the**
+    Don't use.
+<a id="word-list--resource-record-set"></a>
+
+**resource record set**
+    Not *resource recordset*.
+<a id="word-list--retarded"></a>
+
+**Don't use: retarded**
+    Don't use. If you are referring to a system or component being slowed,
+    use the word *slowed*.
+<a id="word-list--retriable"></a>
+
+**Don't use: retriable, triable**
+    Don't use *retriable* or *triable*, unless a code item uses that
+    spelling. Outside of code font, write around the term.
+<a id="word-list--retryable"></a>
+
+**Avoid: retryable, tryable**
+    Where possible, write around *retryable* and *tryable*. For
+    example, write out *you can try it again* or *can be tried
+    again*.
+<a id="word-list--review"></a>
+
+**review**
+    If you mean "read, potentially for the first time," then use *read*
+    instead of *review*.
+    If you mean "read critically, commenting on problems" (as in *code
+    review*), then *review* is fine.
+    Avoid using phrasing like "If you've never heard of OAuth, then review the
+    OAuth documentation."
+<a id="word-list--rfc"></a>
+
+**RFC**
+    When referencing an RFC specification, use a space between *RFC* and
+    the number (for example, *RFC 2318*).
+<a id="word-list--roll-out"></a>
+
+**Avoid: roll out**
+    Don't use to mean a sudden or instantaneous launch. If you use *roll
+    out*, define what you mean. When possible, use a more precise,
+    non-figurative term like *gradual*, *in stages*, *phases*,
+    or *progressive*.
+<a id="word-list--rtfm"></a>
+
+**Don't use: RTFM**
+    Don't use. Instead, use a more precise phrase like "For more information,
+    see ...."
+<a id="word-list--runbook"></a>
+
+**runbook**
+    Not *run book*.
+<a id="word-list--runtime"></a>
+
+**runtime, run time**
+    Use the noun *runtime* when referring to the environment in which
+    software runs, such as a Ruby or Java runtime.
+    Use the noun phrase *run time* when referring to the time during
+    program execution when something occurs, as contrasted with *compile
+    time*, for example.
+    Recommended: The profiler collects
+    data at run time, and the scheduler uses this data at compile time to
+    improve performance for subsequent runs.
+    Recommended: The App Engine standard
+    environment has two generations of runtime environments. The
+    second-generation runtimes significantly improve the capabilities of App
+    Engine.
+
+<a id="word-list--letter-s"></a>
+
+##### S
+
+<a id="word-list--saas"></a>
+
+**SaaS**
+    Write out on first mention: *software as a service (SaaS)*.
+<a id="word-list--sane"></a>
+
+**Don't use: sane**
+    Don't use. Instead use a word like *valid* or *sensible*.
+<a id="word-list--sanity-check"></a>
+
+**Don't use: sanity check**
+    Don't use. Instead, use a term like *quick check*, *confidence
+    check*, *preliminary check* or *coherence check*.
+<a id="word-list--sap"></a>
+
+**SAP**
+    Pronounced as the individual letters *S*, *A*, *P*, so
+    write *an SAP system*, not *a SAP system*. For more information, see
+    [Indefinite articles before abbreviations](#abbreviations--articles).
+<a id="word-list--scale"></a>
+
+**scale**
+    Don't use *scale* alone to say that something is large or increasing.
+    Include supporting words to indicate magnitude or direction of change in
+    magnitude, whether scaling up or down, such as when you change a machine
+    type to add or remove CPUs or RAM, or scaling out or in, such as adding or
+    removing instances from a group.
+    Recommended: The system performs
+    better at a larger scale.
+    Not recommended: The system performs
+    better at scale.
+    Recommended: The system scales up
+    quickly, but it scales down more slowly.
+    Not recommended: The system scales
+    quickly.
+<a id="word-list--screenshot"></a>
+
+**screenshot (noun)**
+    Not *screen shot* or *screensnap*.
+    Don't use as a verb; instead, use *take a screenshot*.
+<a id="word-list--scroll"></a>
+
+**scroll**
+    OK to use *scroll* as a verb, but if possible, instead use a term
+    that isn't specific to implementation. For example, write *go to the
+    section*, instead of *scroll to the section*.
+    If you use *scroll*, don't use directional language
+    like *scroll up*. For more information, see
+    [Accessibility](#accessibility--document-rendering).
+<a id="word-list--search"></a>
+
+**Search (as part of product name)**
+    Capitalize *Search* when referring to a product like Google Search.
+<a id="word-list--search-console"></a>
+
+**Search Console**
+    Capitalize each word in *Search Console*.
+<a id="word-list--see"></a>
+
+**see**
+    OK as a general term and when referring to links and cross-references. Our
+    research indicates that language relating to sight is OK for a wide range
+    of readers. For more information, see
+    [Cross-references and linking](#cross-references).
+<a id="word-list--select"></a>
+
+**select**
+    Use to describe choosing an item from among multiple options, selecting
+    text, or marking a checkbox.
+    Recommended: Select **Automatically
+    check for updates**.
+    Not recommended: Check
+    **Automatically check for updates**.
+<a id="word-list--sensitive"></a>
+
+**sensitive**
+    *Sensitive* data is data for which the release might be harmful. See
+    [confidential](#word-list--confidential).
+<a id="word-list--service"></a>
+
+**service**
+    It's OK to refer to Google products, such as Google Kubernetes Engine or
+    Compute Engine, as *services*. However, if the term *services*
+    leads to ambiguity, then use the product names.
+<a id="word-list--service-level-agreement"></a>
+
+**service level agreement**
+    Lowercase when referring to service level agreements in general.
+    It's OK to use title case (*Service Level Agreement*) when referring
+    to a specific document.
+    OK to abbreviate as *SLA* after first use.
+<a id="word-list--service-level-indicator"></a>
+
+**service level indicator**
+    Lowercase except at the beginning of a sentence,
+    heading, or list item.
+    OK to abbreviate as *SLI* after first use.
+<a id="word-list--service-level-objective"></a>
+
+**service level objective**
+    Lowercase except at the beginning of a sentence,
+    heading, or list item.
+    OK to abbreviate as *SLO* after first use.
+<a id="word-list--setup"></a>
+
+**setup (noun or adjective), set up (verb)**
+<a id="word-list--sexy"></a>
+
+**Don't use: sexy**
+    Don't use. Instead, use precise, positive words, such as *fast*,
+    *powerful*, or *elegant*.
+<a id="word-list--sha-1"></a>
+
+**SHA-1**
+    Not *SHA1*, except in string literals/enums and in hyphenated phrases
+    such as *HSA-SHA1*.
+<a id="word-list--shall"></a>
+
+**Avoid: shall**
+    Avoid *shall* except under advice from a lawyer. For more
+    information, see [should](#word-list--should).
+<a id="word-list--she"></a>
+
+**Avoid: she, her, hers**
+    Don't use a gendered pronoun except for a specific individual of known
+    gender. Use *they* and *their* for the general singular pronoun.
+<a id="word-list--sherpa"></a>
+
+**Don't use: sherpa**
+    If possible, use a more precise term. For example, if you mean
+    *guide*, use that term.
+<a id="word-list--shift-left"></a>
+
+**Avoid: shift left**
+    In general, avoid using this term to mean moving something earlier in
+    time. Instead, use a less figurative phrase, such as *shift earlier*
+    or *move to an earlier phase*. This figurative term relies on the
+    non-universal assumption that the natural flow is from left to right.
+    It's OK to use *shift left* and *shift right* in the context of
+    binary multiplication and division.
+<a id="word-list--should"></a>
+
+**Avoid: should, should be**
+    Generally avoid.
+    Because *should* is ambiguous by definition, it can be problematic. For more information
+    and alternatives, see
+    [Word choice for recommendations and requirements](#prescriptive-documentation--word-choice).
+    See also [can](#word-list--can), [could](#word-list--could),
+    [may](#word-list--may), [might](#word-list--might),
+    [must](#word-list--must), and [would](#word-list--would).
+<a id="word-list--sign-in"></a>
+
+**sign-in (noun or adjective), sign in (verb)**
+    Not *log in* or *signin*.
+<a id="word-list--sign-into"></a>
+
+**Don't use: sign into**
+    Don't use. Instead, use *sign in to*.
+<a id="word-list--sign-on"></a>
+
+**Don't use: sign-on, sign on**
+    Don't use either form on its own. Use the hyphenated version as part of
+    *single sign-on*.
+<a id="word-list--sign-out"></a>
+
+**sign-out (noun or adjective), sign out (verb)**
+    Not *log out* or *signout*.
+<a id="word-list--simple"></a>
+
+**Avoid: simple, simply**
+    What might be simple for you might not be simple for others. Try
+    eliminating this word from the sentence because usually the same meaning
+    can be conveyed without it.
+<a id="word-list--since"></a>
+
+**since**
+    If you mean *because*, then use *because* instead of
+    *since*. *Since* is ambiguous; it can refer to the passage of
+    time. *Because* refers to causation or the reason for something.
+<a id="word-list--single-most"></a>
+
+**single most**
+    Not *singlemost*.
+<a id="word-list--single-pane-of-glass"></a>
+
+**single pane of glass**
+    Avoid. This term is used to favorably compare a centralized control and
+    monitoring interface against the alternative of several disparate
+    interfaces. It can almost always be replaced by *single interface* or
+    *unified interface*.
+<a id="word-list--single-sign-on"></a>
+
+**single sign-on (noun or adjective)**
+<a id="word-list--slave"></a>
+
+**Don't use: slave**
+    Don't use. Instead, use alternative terms appropriate to your domain, such
+    as *worker* or *replica*.
+    If you're replacing the terms *master* and *slave* together,
+    then consider such combinations as *primary*/*secondary*,
+    *primary*/*replica*, *original*/*replica*,
+    *controller*/*worker*, *initiator*/*responder*,
+    *mixer*/*leaf*, *aggregator*/*collector*,
+    *publisher*/*subscriber*, *leader*/*follower*, and
+    *active*/*standby*.
+    If the command or code that you're documenting uses the literal word
+    *slave*, then use this word only in direct reference to the code item
+    ([formatted as code](#code-in-text)), make it clear what
+    you're referring to, and use the new term thereafter. For example, "Invoke
+    the secondary (`slave`) process directly when debugging issues
+    between the primary and secondary processes."
+    See also [master](#word-list--master).
+<a id="word-list--slice"></a>
+
+**Don't use: slice and dice**
+    Don't use the phrase *slice and dice*. Instead, use specific terms
+    appropriate to the task that you're describing. Some possible options
+    include: *segment data for analysis* or *break information into
+    smaller parts*.
+<a id="word-list--smartphone"></a>
+
+**Don't use: smartphone, smart phone**
+    Don't use. Instead, use [*mobile phone*](#word-list--mobile) or
+    *phone*. If you're talking about more than phones, then use *mobile
+    device*. It's OK to use *phone* (without *mobile*) when the
+    context is clear.
+<a id="word-list--soon"></a>
+
+**Avoid: soon**
+    Avoid in timeless documentation because this word can become outdated. The
+    word can also prematurely disclose product or feature strategy or
+    inappropriately imply that a product or feature might change.
+    See also [eventually](#word-list--eventually) and
+    [future](#word-list--future).
+    Recommended: This setting is
+    optional.
+    Not recommended: This setting is
+    optional for existing applications but will soon be required for all
+    applications.
+    For more information, see
+    [Timeless documentation](#timeless-documentation).
+<a id="word-list--spin-up"></a>
+
+**Avoid: spin up**
+    As in *spin up an instance*. Avoid using *spin up* unless you're
+    referring to a hard disk; instead, use a less colloquial term like
+    *create* or *start*.
+<a id="word-list--sql"></a>
+
+**SQL**
+    Refer to *a SQL* (pronounced "a sequel"), not *an SQL*. For more
+    information, see
+    [Indefinite articles before abbreviations](#abbreviations--articles).
+<a id="word-list--ssh"></a>
+
+**ssh and SSH**
+    Don't use `ssh` or SSH as a verb. SSH is a secure
+    communications protocol; `ssh` is a utility.
+    Recommended: To establish an SSH
+    connection, use the `ssh` command.
+    Recommended: Connect to the instance
+    by using SSH.
+        Not recommended: `ssh` into
+        your remote shell.
+<a id="word-list--sshing"></a>
+
+**Don't use: ssh'ing**
+    Don't use. See also [ssh and SSH](#word-list--ssh).
+    Recommended: When you use
+    `ssh` to log in ...
+<a id="word-list--startup"></a>
+
+**startup (noun or adjective), start up (verb)**
+<a id="word-list--static-external-ip-address"></a>
+
+**static external IP address**
+    Don't use *static IP address* or *external IP address* to refer
+    to static external IP addresses.
+<a id="word-list--status-bar"></a>
+
+**status bar**
+    Not *statusbar* or *status-bar*.
+    Lowercase except at the beginning of a sentence,
+    heading, or list item.
+<a id="word-list--stonith"></a>
+
+**Avoid: STONITH, STOMITH**
+    Avoid using
+    [graphic or
+    metaphorical language](#inclusive-documentation--graphic-language). Instead, explain the relevant feature, such as
+    *fence failed nodes*.
+<a id="word-list--style-sheet"></a>
+
+**style sheet**
+    *Style sheet* and *stylesheet* are both acceptable spellings. However, be consistent
+    with your choice throughout a given document.
+<a id="word-list--sub-command"></a>
+
+**sub-command**
+    Not *subcommand*.
+<a id="word-list--subnet"></a>
+
+**subnet**
+    OK to use as a shortening of *subnetwork*. Use the same term consistently throughout your
+    document. For more
+    information, see [Subnets vs. subnetworks](https://cloud.google.com/compute/docs/vpc/#subnets_vs_subnetworks).
+<a id="word-list--subtree"></a>
+
+**subtree**
+    Not *sub-tree*.
+<a id="word-list--subzone"></a>
+
+**subzone**
+    Not *sub-zone* or *sub zone*.
+<a id="word-list--such-as"></a>
+
+**such as**
+    Use *such as* to introduce examples or draw comparisons. Note that *such as*,
+    *like*, and *include* introduce non-exhaustive lists, so it's redundant to combine
+    them with *etc.*, *so forth*, or *and more*. See also [etc.](#word-list--etc),
+    [like](#word-list--like). For more information, see [Format
+    examples](#format-examples).
+<a id="word-list--surface"></a>
+
+**surface**
+    Avoid as a transitive verb; instead, use a more specific term, such as
+    *make people aware of* or *expose*.
+    Recommended: To make the audit logs
+    available, you must configure the monitoring system.
+    Not recommended: To surface audit
+    logs, you must configure the monitoring system.
+
+<a id="word-list--letter-t"></a>
+
+##### T
+
+<a id="word-list--tab"></a>
+
+**tab**
+    When referring to the sub-pages of a [console](#word-list--console), use
+    *page* instead of *tab*.
+<a id="word-list--table-name"></a>
+
+**table name**
+    Two words. Set specific table names in code font.
+<a id="word-list--tablet"></a>
+
+**tablet**
+    *Tablet* is OK. If you don't know whether it's a tablet or a phone,
+    use *device*.
+<a id="word-list--tag"></a>
+
+**tag**
+    See [element](#word-list--element).
+<a id="word-list--tap"></a>
+
+**(Android) tap**
+    In Android documentation, use for on-screen and soft (capacitive)
+    buttons.
+        Use instead of *click* when the environment is definitely a
+        touch device.
+        Use instead of *touch*. However, *touch & hold* (not *touch
+        and hold*) is OK to use.
+        For mechanical buttons, use [*press*](#word-list--press).
+<a id="word-list--tap-and-hold"></a>
+
+**Don't use: (Android) tap & hold, tap and hold**
+    In Android documentation, don't use. Instead, use *touch & hold*.
+    (Not *touch and hold*.)
+<a id="word-list--tarball"></a>
+
+**Don't use: tarball**
+    Don't use. Instead, use *tar file*.
+<a id="word-list--target"></a>
+
+**target**
+    Avoid using as a verb when possible, especially in reference to people.
+    For some readers, *target* has aggressive connotations. Instead of
+    "targeting" audiences, we try to attract them or appeal to them or make
+    their lives easier.
+    It's OK to use *target* as an adjective, as in *target
+    audience*, but consider rephrasing for clarity. Alternatives
+    include phrases such as *intended for*, *looking for*,
+    *focused on*, and *interacting with*.
+<a id="word-list--terminate"></a>
+
+**Avoid: terminate**
+    Avoid using as a synonym for *stop*. Instead, use words like
+    *stop*, *exit*, *cancel*, or *end*.
+    For a specific context where you can use *terminate* as a synonym for
+    *stop*, see [Documenting
+    command-line syntax](#code-syntax--linux-signals).
+        In some contexts, such as telephony and networking, *terminate* has
+            specific technical meanings that aren't synonyms for *stop*; in those
+            contexts, you can use *terminate*.
+<a id="word-list--textbox"></a>
+
+**Don't use: text box, textbox**
+    Don't use. Instead, use *box*. For more information, see
+    [Text box](#ui-elements--term-textbox).
+    In Google Cloud documentation, use
+    *field* instead of *box*. For example, "In the **Instance**
+    field, specify a value less than 64 characters long."
+    In Google Workspace documentation, use
+    *field* instead of *box*. For example, "In the **Instance**
+    field, specify a value less than 64 characters long."
+<a id="word-list--their"></a>
+
+**their (singular)**
+    See [*they*](#word-list--they).
+<a id="word-list--then"></a>
+
+**then**
+    Although it is common in casual usage to omit the word *then* in *if...then*
+    statements, you should include helper words like *then* in technical documentation. For
+    more information, see
+    [Use clear, precise, and unambiguous language](https://developers.google.com/style/translation#clear-language).
+<a id="word-list--they"></a>
+
+**they (singular)**
+    This is our preferred gender-neutral pronoun. Whether used as singular
+    or plural, it always takes the plural verb. For example, "A user
+    authenticates their identity by entering their password." See also [gender-neutral he](#word-list--gender).
+<a id="word-list--third-party"></a>
+
+**third party (noun), third-party (adjective)**
+    Spell as *third party* when used as a noun and as *third-party* when used as an
+    adjective.
+    Avoid abbreviating to *3rd party* or *3rd-party*. For more information, see [Ordinal numbers](#numbers--ordinal-numbers).
+<a id="word-list--this-that"></a>
+
+**this, that**
+    Where possible, put a noun after *this* or *that* for clarity.
+    If doing so results in clunky prose, then don't do it; but even then, try
+    thinking about what the noun would be. If you aren't sure what noun
+    *this* or *that* refers to, then consider rephrasing—otherwise, your reader
+    probably won't know what noun you're referring to, either.
+<a id="word-list--time-frame"></a>
+
+**timeframe**
+    Not *time frame*. Avoid where possible, or use an alternative such as
+    *period*, *schedule*, *deadline*, or *when*. But if
+    you do use it, then write it as one word.
+<a id="word-list--timeout"></a>
+
+**timeout (noun), time out (verb)**
+<a id="word-list--time-stamp"></a>
+
+**timestamp**
+    Not *time stamp*.
+<a id="word-list--ttl"></a>
+
+**time to live**
+    Not *time-to-live*. Abbreviate as *TTL* after first use.
+<a id="word-list--time-zone"></a>
+
+**time zone (noun), time-zone (adjective)**
+    Spell as *time zone* when used as a noun and as *time-zone* when used as an
+    adjective. See also [wake lock (noun), wake-lock (adjective)](#word-list--wake-lock).
+<a id="word-list--tldr"></a>
+
+**Don't use: tl;dr**
+    Don't use. Instead, use something like *To summarize*, or revise the
+    sentence.
+<a id="word-list--toolkit"></a>
+
+**toolkit**
+    Not *tool-kit* or *tool kit*.
+<a id="word-list--touch"></a>
+
+**Don't use: (Android) touch**
+    In Android documentation, don't use. Instead, use *tap*. However,
+    *touch & hold* is OK to use.
+<a id="word-list--touch-and-hold"></a>
+
+**(Android) "touch & hold"**
+    Not *touch and hold*.
+<a id="word-list--touchscreen"></a>
+
+**touchscreen**
+    Not *touch screen*
+<a id="word-list--traditional"></a>
+
+**Avoid: traditional**
+    If possible, use a more precise term.
+    Recommended: Conventionally, Python
+    function names are lowercase, with words separated by underscores.
+    Not recommended: Traditionally, Python
+    function names are lowercase, with words separated by underscores.
+    Recommended: This tutorial explains
+    how to migrate from an on-premises data warehouse to BigQuery.
+    Not recommended: This tutorial
+    explains how to migrate from a traditional data warehouse to BigQuery.
+<a id="word-list--transpile"></a>
+
+**transpile**
+    Not *transcompile*.
+<a id="word-list--tribal-knowledge"></a>
+
+**Don't use: tribal knowledge, tribal wisdom**
+    Don't use. Instead, use a less figurative term to indicate knowledge held
+    by a group of people.
+<a id="word-list--trojan"></a>
+
+**trojan**
+    Lowercase when referring to malware.
+<a id="word-list--turn-on"></a>
+
+**Avoid: turn on**
+    In procedures, use the appropriate label and action for the
+    [UI element](#ui-elements) that the user interacts with.
+    For turning on or activating an option or feature, use *turn on* or
+    [enable](#word-list--enable) consistently. Use the same term consistently throughout your
+    document.
+    Recommended: To turn on Magic Mode,
+    follow these steps.
+    Recommended: In **Settings**, click
+    the **Magic mode** toggle to the on position.
+<a id="word-list--tutorial"></a>
+
+**tutorial**
+    OK to use. See [documentation](#word-list--documentation).
+<a id="word-list--type"></a>
+
+**Avoid: type**
+    In general, use [enter](#word-list--enter) instead of *type* because
+    there is typically more than one way to enter text than typing (such as
+    pasting text or speaking).
+<a id="word-list--typically"></a>
+
+**typically**
+    Use to describe what is usual or expected under normal circumstances.
+    Don't use as the first word in a sentence, as doing so can leave the
+    meaning open to misinterpretation.
+
+<a id="word-list--letter-u"></a>
+
+##### U
+
+<a id="word-list--ui"></a>
+
+**UI**
+    Don't use generically to refer to a page or dashboard. Use a more specific
+    term like [*page*](#word-list--page) or
+    [*console*](#word-list--console). If a specific term is unavailable,
+    use *web interface*.
+    Recommended: In the Google Cloud
+    console
+    Recommended: On the **Cloud Tasks**
+    page
+    Recommended: In the Secure Source
+    Manager web interface
+    Not recommended: In the **Cloud
+    Tasks** UI
+<a id="word-list--unarchive"></a>
+
+**Don't use: unarchive**
+    Don't use. Instead, use *extract*.
+<a id="word-list--uncheck"></a>
+
+**Don't use: uncheck**
+    Don't use to refer to clearing a check mark from a checkbox. Instead, use
+    *clear*.
+    Recommended: Clear **Automatically
+    check for updates**.
+    Not recommended: Uncheck
+    **Automatically check for updates**.
+    Not recommended: Deselect
+    **Automatically check for updates**.
+<a id="word-list--uncompress"></a>
+
+**Don't use: uncompress**
+    Don't use. Instead, use *extract*.
+<a id="word-list--under"></a>
+
+**Don't use: under**
+    Don't use for a range of version numbers. Instead,
+    use [*earlier*](#word-list--earlier).
+    Don't use to refer to a position in the UI.
+    Recommended: In the **Service account
+    ID** field, enter a name.
+    Recommended: For **Service account
+    ID**, enter a name.
+    Not recommended: Under **Service
+    account ID**, enter a name.
+<a id="word-list--unicode"></a>
+
+**Unicode**
+    Not *UNICODE*.
+<a id="word-list--unix-like"></a>
+
+**Unix-like**
+    Not *Unixlike* or *Unix like*.
+<a id="word-list--unix-epoch-time"></a>
+
+**Unix epoch time**
+    Use instead of *Unix time* or *epoch time* to refer to a
+    point in time represented as a number of seconds since the Unix epoch
+    (00:00:00 UTC on January 1, 1970), ignoring leap seconds.
+<a id="word-list--unselect"></a>
+
+**Don't use: unselect**
+    Don't use. Instead, use *clear* for checkboxes, and *deselect*
+    for other UI elements.
+<a id="word-list--unsighted"></a>
+
+**Don't use: unsighted**
+    Don't use. See [blind](#word-list--blind).
+<a id="word-list--untar"></a>
+
+**Don't use: untar**
+    Don't use. Instead, use *extract*.
+<a id="word-list--unzip"></a>
+
+**Don't use: unzip**
+    Don't use. Instead, use *extract*.
+<a id="word-list--us"></a>
+
+**US**
+    OK to use as an abbreviation for *United States*. Don't use
+    *U.S.* or *U.S.A.* For more information, see [Periods with abbreviations](#abbreviations--periods).
+<a id="word-list--user"></a>
+
+**user**
+    Use the word *user* only to refer to the user of the software that
+    your reader is developing. Otherwise, address the reader as *you*
+    and assume that they will complete the tasks that you're documenting. For
+    more information, see [Second person and first
+    person](#person).
+<a id="word-list--user-base"></a>
+
+**user base**
+    Not *userbase*.
+<a id="word-list--using"></a>
+
+**using**
+    Where *using* might have more than one interpretation, use *by
+    using* to help clarify the logic of the sentence.
+    Recommended: You can filter for data
+    with specific attributes by using custom filters.
+    Not recommended: You can filter for
+    data with specific attributes using custom filters.
+<a id="word-list--utf"></a>
+
+**UTF**
+    Include the hyphen in the names of Unicode encodings, such as
+    *UTF-8*, *UTF-16*, and *UTF-32*.
+<a id="word-list--utilize"></a>
+
+**utilize, utilization**
+    Use with caution. Don't use *utilize* when you mean *use*. It's
+    OK to use *utilize* or *utilization* when referring to the
+    quantity of a resource being used.
+    Recommended: When CPU utilization
+    exceeds 75%, the autoscaler adds more CPU resources.
+    Recommended: To distribute network
+    traffic, use a load balancer.
+    Not recommended: To distribute network
+    traffic, utilize a load balancer.
+
+<a id="word-list--letter-v"></a>
+
+##### V
+
+<a id="word-list--v"></a>
+
+**v (abbreviating *version*)**
+    Use lowercase.
+<a id="word-list--via"></a>
+
+**Don't use: via**
+    Don't use.
+<a id="word-list--vice-versa"></a>
+
+**Don't use: vice versa**
+    Don't use. Instead, use a phrase like *the other way around*,
+    *conversely*, or *otherwise*. In some contexts, vice versa is
+    unclear or imprecise because in a complex sentence it's hard to know which
+    two things are swapped with each other. In such cases, make it explicitly
+    clear what two things are swapped.
+<a id="word-list--virtual-machine-instance"></a>
+
+**virtual machine (VM) instance**
+    Use when first introducing virtual machines on a given page. For
+    subsequent mentions, you can use *VM instance* or *VM*.
+    For Google Cloud: on first mention of a Compute Engine VM,
+    use *Compute Engine instance* and then use *compute instance*
+    throughout the rest of the document. If you need to indicate other types of
+    VMs, use *VM*, *VM instance*, or *bare metal instance*.
+    See also [GKE node](#word-list--gke-node).
+<a id="word-list--visually-challenged"></a>
+
+**Don't use: visually challenged**
+    See [blind](#word-list--blind).
+<a id="word-list--vlan"></a>
+
+**VLAN attachment**
+    Don't use the following: *interconnect attachment (VLAN)*,
+    *Interconnect attachment*, *Cloud Interconnect attachment*, or
+    any variation thereof. See also
+    [interconnectAttachment](#word-list--interconnect-attachment).
+<a id="word-list--voila"></a>
+
+**Don't use: voila**
+    Don't use.
+<a id="word-list--voodoo"></a>
+
+**Don't use: voodoo**
+    Don't use. Instead, use a term like *mysterious*, *complicated*,
+    or *nondeterministic*.
+<a id="word-list--vs"></a>
+
+**Don't use: vs.**
+    Don't use *vs.* as an abbreviation for *versus*; instead, use
+    the unabbreviated *versus*.
+
+<a id="word-list--letter-w"></a>
+
+##### W
+
+<a id="word-list--wake-lock"></a>
+
+**wake lock (noun), wake-lock (adjective)**
+    Spell as *wake lock* when used as a noun and as *wake-lock* when used as an
+    adjective. See also [time zone (noun), time-zone (adjective)](#word-list--time-zone).
+<a id="word-list--walkthrough"></a>
+
+**walkthrough**
+    Not *walk-through*.
+<a id="word-list--war-room"></a>
+
+**Don't use: war room, warroom, war-room**
+    Don't use. Instead, use a more precise term to describe the activity or
+    team. Depending on context, possible alternatives include *rapid
+    response team*, *situation response team*, *situation room*,
+    *incident-management team*, or *media monitoring room*.
+<a id="word-list--warm"></a>
+
+**warm**
+    When possible, avoid [jargon](#jargon) like *warm
+    failover*, *warm standby*, and *warm spare*. If you use one
+    of these phrases, define it on first use and use it consistently
+    throughout the document.
+<a id="word-list--we"></a>
+
+**we**
+    Don't use *we* (or other first-person plural pronouns such as
+    *our* or *us*) to address the reader who is performing the
+    tasks that you're documenting. Instead, use *you*.
+    It's OK to use *we* to refer to the organization that's represented
+    as the author of the document as long as the antecedent is clear. For more
+    information, see
+    [Second person and first person](#person).
+<a id="word-list--web"></a>
+
+**web (lowercase)**
+<a id="word-list--wasm"></a>
+
+**WebAssembly, Wasm**
+    Use the capitalization established in the
+    [WebAssembly specification](https://webassembly.github.io/spec/core/intro/introduction.html#introduction).
+<a id="word-list--web-application-firewall"></a>
+
+**web application firewall (lowercase)**
+<a id="word-list--webmaster"></a>
+
+**Don't use: webmaster, web master**
+    Don't use. Instead, use a more precise term to describe the specific role,
+    such as *website owner*, *website administrator*, *web content
+    manager*, *owner of a site*.
+<a id="word-list--web-server"></a>
+
+**web server**
+    Not *webserver*.
+<a id="word-list--whether"></a>
+
+**whether**
+    - To decide whether it's more appropriate to use *if* or
+      *whether*, see [Grammar Girl's
+      discussion of *if* and *whether*](http://www.quickanddirtytips.com/education/grammar/if-versus-whether).
+    - To decide whether you need to add *or not* when using
+      *whether*, see [the New York
+      Times's blog post about whether (or not)](http://afterdeadline.blogs.nytimes.com/2010/03/01/whether-or-not/).
+<a id="word-list--while"></a>
+
+**while**
+    Don't use to indicate a contrast. Instead, use a more precise term, such
+    as *although*.
+    OK to use to refer to a period of time.
+<a id="word-list--white-box"></a>
+
+**Avoid: white-box**
+    Avoid using *white-box*, *whitebox*, or *white box* to
+    describe monitoring and testing. Consider using a more precise term for
+    clarity.
+
+    - For monitoring, use *introspective monitoring*.
+    - For testing, use *clear-box testing*.
+<a id="word-list--white-glove"></a>
+
+**Avoid: white glove, white-glove, whiteglove**
+    Avoid using. Instead use terms like *high-touch*, *premium*, or
+    *platinum-level*.
+<a id="word-list--whitehat"></a>
+
+**Don't use: whitehat, white hat, white-hat**
+    Don't use. Instead, use precise terms for the kind of compliance, such as
+    *legal*, *ethical*, or *following the rules*.
+<a id="word-list--white-label"></a>
+
+**Avoid: white label, whitelabel, white-label**
+    Don't use. Instead, use a more precise term for your context, such as
+    *unbranded*, *unlabeled*, or *blank label*.
+<a id="word-list--whitelist"></a>
+
+**Don't use: whitelist, white list, white-list**
+    Don't use. See [blacklist](#word-list--blacklist).
+<a id="word-list--whitelisted"></a>
+
+**Don't use: whitelisted, white listed, white-listed**
+    Don't use. See [blacklist](#word-list--blacklist).
+<a id="word-list--whitelisting"></a>
+
+**Don't use: whitelisting, white listing, white-listing**
+    Don't use. See [blacklist](#word-list--blacklist).
+<a id="word-list--whitepaper"></a>
+
+**whitepaper**
+    Not *white paper*.
+    When possible, use a more precise term. The term *whitepaper* has a variety of
+    meanings in various contexts. If you must use the term *whitepaper*, also use descriptive
+    terms to provide context.
+<a id="word-list--whitespace"></a>
+
+**whitespace**
+    Not *white space*.
+<a id="word-list--wildcard"></a>
+
+**wildcard**
+    Not *wild card*.
+<a id="word-list--will"></a>
+
+**will**
+    Avoid. Applies equally to its past tense, *would*. See also
+    [Present tense](#tense) and
+    [Documenting future features](#future).
+<a id="word-list--wish"></a>
+
+**Don't use: wish**
+    Don't use. Instead, use a word like *want* or *need*.
+<a id="word-list--with"></a>
+
+**with**
+    Don't use *with* when expressing ownership:
+        Recommended: A handset that has 2 GB
+        of RAM.
+        Not recommended: A handset with 2 GB
+        of RAM.
+    Don't use *with* when expressing use:
+        Recommended: Use the debugging tool
+        to debug.
+        Not recommended: Debug this tool with
+        the debugging tool.
+<a id="word-list--workload"></a>
+
+**workload**
+    The term *workload* might refer to software, like an app or
+    a service; to app resources, like data and infrastructure; or to physical
+    components that work together.
+    Where possible, use a more precise term to describe what you mean. If you
+    use the term *workload*, define your meaning on first use as you
+    normally would with jargon and other ambiguous terms.
+<a id="word-list--world-wide-web"></a>
+
+**Don't use: World Wide Web**
+    Don't use. Instead, use *web*.
+<a id="word-list--would"></a>
+
+**Avoid: would**
+    Avoid using. Instead, use *can* where possible.
+    See also [can](#word-list--can), [could](#word-list--could),
+    [may](#word-list--may), [might](#word-list--might),
+    [must](#word-list--must), and [should](#word-list--should).
+    For information about clarifying who's performing an action, see
+    [Active voice](#voice).
+    For information about tenses, see [Present
+    tense](#tense).
+
+<a id="word-list--letter-y"></a>
+
+##### Y
+
+<a id="word-list--ymmv"></a>
+
+**Don't use: ymmv**
+    Don't use. Instead, use something like *Your results might vary*.
+<a id="word-list--you"></a>
+
+**you**
+    Use *you* instead of [*user*](#word-list--user) to address the
+    reader of your document. For more information, see
+    [Second person and first person](#person).
+
+<a id="word-list--letter-z"></a>
+
+##### Z
+
+<a id="word-list--zippy"></a>
+
+**Avoid: zippy**
+    Don't use to refer to [expander arrows](#word-list--expander-arrow),
+    unless you're specifically referring to the [Zippy widget](https://google.github.io/closure-library/api/goog.ui.Zippy.html)
+    in Closure.
+
+---
+
+<a id="product-names"></a>
+
+### Product names
+
+*Source: <https://developers.google.com/style/product-names>*
+
+This page describes how to use product names.
+
+<a id="product-names--capitalize"></a>
+
+#### Capitalize product names
+
+In general, Google product names are in *title case*, sometimes called
+*init-capped*, which means that every word is capitalized except for
+prepositions like *of* or *on* and articles like *a* or
+*the*. When you refer to a Google product, use title case except
+when you're matching a UI label. For information about how to refer to UI
+labels, see
+[UI elements and interaction](#ui-elements).
+
+When you write about any product, follow the official capitalization for the
+names of brands, companies, software, products, services, features, and
+terms defined by companies and open source communities.
+
+- For example, if you're using Kubernetes-related terms, then follow
+  the capitalization that's shown in the Kubernetes [Concepts
+  documentation](https://kubernetes.io/docs/concepts/).
+
+  Recommended in a Kubernetes
+  context: A Job creates one or more Pods.
+
+  Recommended: The Cloud Scheduler
+  job publishes a message to a Pub/Sub topic at one-minute intervals.
+- If an official name begins with a lowercase letter, then put it in
+  lowercase even at the start of a sentence. But it's better to revise
+  the sentence to avoid putting a lowercase word at the start, if
+  possible.
+
+  Recommended: You can use macOS to
+  run the app.
+
+  Not recommended: macOS can run the
+  app.
+
+<a id="product-names--feature-names"></a>
+
+##### Feature names
+
+A *feature* is a distinctive attribute or capability of a product.
+Features are usually described in terms of what they can do as part of a
+product. In general, feature names are lowercase, although there are
+exceptions.
+
+When you write about a feature, don't capitalize it unless the name is
+officially capitalized. If you're unsure, follow the precedent that's set
+by other documents that describe the feature. As with products, match
+the capitalization of a UI label if you're referring to one.
+
+For more general information about capitalization, see
+[Capitalization](#capitalization).
+
+<a id="product-names--shortening"></a>
+
+#### Shorten Google product names
+
+When referring to a Google product, sometimes you might want to abbreviate
+the product name. For example, when you're referring to Google
+Spreadsheets, it can be awkward to refer to it as Google Spreadsheets
+every time; sometimes you might want to call it Spreadsheets.
+
+Use the full trademarked product name. Don't abbreviate product names,
+except in cases where you're matching a UI label. In such cases, make it
+clear that you're referring to the Google product and not some other thing
+with a similar name.
+
+Also consider whether you need to refer to a product name throughout a
+document, or if you can use a more general term. For example, if you've
+established that you're talking about *Anthos Service Mesh*, you can
+probably frame your discussion around the concept of *a service mesh*
+throughout much of the document.
+
+<a id="product-names--possessives"></a>
+
+<a id="product-names--possessives-of-product-names"></a>
+
+#### Possessives of product names
+
+For information about forming possessives with product names, see
+[Product, feature, and company names](#possessives--productu002c-featureu002c-and-company-names).
+
+<a id="product-names--the-with-names"></a>
+
+<a id="product-names--articles-before-product-names"></a>
+
+#### Articles before product names
+
+Don't use *the* before a product name unless you're using the name to
+modify something else. *Do* use *the* before tool and API names.
+
+Recommended: Using Cloud Datastore with Cloud Dataproc
+
+Recommended: The Cloud Datastore options page
+
+Recommended: The Google Cloud console
+
+Recommended: The Transcoder API
+
+Recommended: The `gcloud` CLI
+
+Not recommended: Using the Cloud Datastore with Cloud
+Dataproc
+
+If you use a product name as a modifier with an indefinite article (*a* or *an*), pay
+close attention to which article precedes the product name.
+
+Recommended: An Anthos Service Mesh environment
+
+Recommended: A Service Mesh environment
+
+For more information about using articles, see [Articles](#articles).
+
+<a id="product-names--multiple-products"></a>
+
+<a id="product-names--use-service-to-refer-to-multiple-products"></a>
+
+#### Use "service" to refer to multiple products
+
+It's OK to refer to Google products as services, such as *the Google Kubernetes Engine
+service* or *the Compute Engine service*. However, if the term *services* leads to
+ambiguity, use the product names.
+
+<a id="product-names--product-names-as-verbs"></a>
+
+<a id="product-names--dont-use-product-names-as-verbs"></a>
+
+#### Don't use product names as verbs
+
+Don't use product names or feature names as verbs.
+
+---
+
+<a id="text-formatting"></a>
+
+### Text-formatting summary
+
+*Source: <https://developers.google.com/style/text-formatting>*
+
+The page summarizes, and provides a quick reference for, many of the general text-formatting
+conventions covered elsewhere in the style guide. For more information, see
+[Visual formatting](#semantic-tagging--visual-formatting).
+
+<a id="text-formatting--bold"></a>
+
+**Bold**
+    Use bold formatting, `<b>` or `**`, only for
+    [UI elements](#ui-elements--formatting) and
+    [run-in headings](#lists--types-of-lists), including at the beginning of
+    [notices](#notices).
+    Although a double underscore, `__`, can also indicate bold styling in Markdown, it
+    can be difficult to distinguish in a text editor. It's best to use the double asterisk for bold in
+    Markdown.
+<a id="text-formatting--italic"></a>
+
+**Italic**
+    In general, use italics sparingly.
+    When you're discussing or introducing terms, such as when defining terms or using
+    *words as words*, use italics formatting, `<i>` or `_`. For more
+    information, see
+    [Use italics to discuss terms](#italics-terms)
+    and
+    [Format abbreviation introductions](#abbreviations--format-abbreviations).
+    When you need to add emphasis to indicate importance, use italics, not bold or underline. But
+    usually, your words can carry the emphasis without adding italics. To indicate
+    [semantic emphasis](#semantic-tagging) in HTML, use the `em` element,
+    which renders as italics in most contexts. To indicate emphasis in Markdown, use underscores
+    (`_`), which render as italics; you can't do semantic tagging in Markdown.
+    Although an asterisk, `*`, can also indicate italics in Markdown, we recommend
+    underscores to make it easier for humans to distinguish italics from bold in the Markdown file.
+    Italicize titles of books, movies, web series, and other full-length works, unless they're part
+    of a link. For more information, see
+    [Cross-references and linking](#cross-references).
+    Italicize mathematical variables—for example, *x* + *y* = 3.
+    Don't italicize mathematical operators such as the plus sign. For more information about
+    formatting mathematical notation, see
+    [Mathematical notation](#mathematical-notation--format-mathematical-notation).
+    Italicize version variables—for example, version 1.4.*x*.
+<a id="text-formatting--underline"></a>
+
+**Underline**
+    Reserve underlining for link text. For more information, see
+    [Style link text](https://developers.google.com/style/cross-references#style-link-text).
+<a id="text-formatting--code-font"></a>
+
+**Code font**
+    Use `<code>` in HTML or `` ` `` in Markdown to apply a monospace font
+    and other styling to [code in text](#code-in-text), inline code, and user
+    input.
+    Use code blocks, `<pre>` or `` \`\`\` ``, for
+    [code samples](#code-samples) or other blocks of code.
+    Do not override or modify font styles inline.
+    Use code font to mark up code, such as filenames, class names, method names, HTTP status codes,
+    console output, and placeholders. For more information, see
+    [Some specific items to put
+    in code font](#code-in-text--some-specific-items-to-put-in-code-font).
+<a id="text-formatting--capitalization"></a>
+
+**Capitalization**
+    Use American English style for
+    [general capitalization](#capitalization).
+    Use sentence case in all [headings,
+    titles, and navigation](#capitalization--capitalization-in-titles-and-headings).
+    Use all-capitals for [placeholders](#placeholders--placeholder-text).
+<a id="text-formatting--quotation-marks"></a>
+
+**Quotation marks**
+    In general, use American English style when [punctuating
+    quotations](#quotation-marks).
+    For titles of shorter works—such as articles or episodes in a web series—put titles in quotation
+    marks, unless they're part of a link.
+<a id="text-formatting--font-style"></a>
+
+**Font type, size, and color**
+    Do not override global styles for [font type, size, or
+    color](https://developers.google.com/style/fonts).
+    Use [semantic HTML](#semantic-tagging) or Markdown to
+    control the style of text on a page—for example, code tags in HTML (`<code>`)
+    or backticks in Markdown (`` ` ``)—instead of manually styling text with a monospace
+    font.
+<a id="text-formatting--other-punctuation"></a>
+
+**Other punctuation conventions**
+    Don't use [ampersands (&)](#word-list--ampersand) as conjunctions or
+    shorthand for *and*. Use *and* instead. That includes headings and navigation.
+    **Exception**: It's okay to use *&* in cases where you need to refer to a UI
+    element or the name of a menu that uses *&*.
+    Put quotation marks and end punctuation outside of link text. For more information, see
+    the [Punctuation around link text](https://developers.google.com/style/cross-references#punctuation)
+    and [Quotation marks and italics](https://developers.google.com/style/cross-references#quotation-marks-italics)
+    sections of the "Cross-references and linking" page.
+
+<a id="text-formatting--more-resources"></a>
+
+#### More resources
+
+- [Mathematical notation](#mathematical-notation)
+
+---
+
+## General principles
+
+<a id="accessibility"></a>
+
+### Accessibility
+
+*Source: <https://developers.google.com/style/accessibility>*
+
+We write our developer documentation with accessibility in mind. This page is not an exhaustive
+reference, but describes some general guidelines and examples that illustrate best practices to
+follow. The
+[World Health Organization](https://www.who.int/en/news-room/fact-sheets/detail/disability-and-health)
+estimates that 15% of the world's population (more than 1 billion people) have an accessibility
+need. When documentation is written with accessibility in mind, it improves the overall
+experience for all readers.
+
+For other writing best practices, see the following resources:
+
+- [Write for a global audience](#translation)
+- [Write inclusive documentation](#inclusive-documentation)
+- [Voice and tone](#tone)
+
+<a id="accessibility--general-dos-and-donts"></a>
+
+#### General dos and don'ts
+
+- Don't use ableist language. Avoid bias and harm when discussing disability and accessibility.
+  For more information, see
+  [Writing inclusive documentation](#inclusive-documentation).
+- Ensure that readers can reach all parts of the document (including
+  tabs, form-submission buttons, and interactive elements) by using only a keyboard,
+  without a mouse or trackpad.
+- Use a screen reader to test your documentation. This test can help you find accessibility
+  issues in your content and is a good way to self-edit your content. To try out a screen reader,
+  see [List of screen readers](https://wikipedia.org/wiki/List_of_screen_readers).
+- In HTML, use [semantic
+  tagging](#semantic-tagging). For example, use the `em` element only to
+  indicate emphasis, not to indicate italics.
+- In HTML, prefer
+  [native
+  elements](https://developer.mozilla.org/en-US/docs/Web/HTML/Element) over custom styles.
+- Avoid unnecessary font formatting. (Screen readers explicitly describe
+  text modifications.)
+- If you're documenting a product that includes specialized accessibility
+  features, then explicitly document those features. For example, the Google Cloud
+  CLI (`gcloud` CLI) includes togglable accessibility features
+  such as percentage progress bars and ASCII box rendering.
+- Don't force line breaks (hard returns) within sentences and paragraphs. Line breaks might not
+  work well in resized windows or with enlarged text.
+- Avoid when possible [camel case](https://wikipedia.org/wiki/Camel_case) and
+  [all caps](https://wikipedia.org/wiki/All_caps). Some screen readers read
+  capitalized letters individually, and some languages are
+  [unicase](https://wikipedia.org/wiki/Unicase). Follow
+  [capitalization](#capitalization) guidelines.
+- Depending on the screen reader (or personal settings), not all punctuation marks are read. Make
+  sure that the same meaning is conveyed to the reader without punctuation marks. For that reason, avoid
+  when possible the use of exclamation marks, question marks, and semicolons.
+- Don't use *&* instead of *and* in headings, text, navigation, or
+  tables of contents. However, it's OK to use *&* when referencing UI
+  elements that use *&*, or in table headings and diagram labels where space
+  constraints require abbreviation. Of course, it's fine to use `&`
+  for technical purposes in code.
+
+<a id="accessibility--ease-of-reading"></a>
+
+#### Ease of reading
+
+- Break up walls of text to aid in scannability. For example, separate
+  [paragraphs](#paragraph-structure),
+  create
+  [headings](#headings),
+  and use
+  [lists](#lists).
+- Use shorter sentences. Try to use fewer than 26 words per sentence.
+- Define acronyms and abbreviations on first usage and if they're used infrequently.
+- Use parallel writing structures for similar things. For example, start each list in the same
+  format.
+- Place distinguishing and important information of a paragraph in the first sentence to aid in
+  scannability.
+- Use clear and direct language. Avoid the use of double negatives and exceptions for exceptions.
+
+  Recommended: You can continue without a
+  path.
+
+  Not recommended: A missing path won't
+  prevent you from continuing.
+- Left-align text for readability. Don't center or full-justify text.
+
+<a id="accessibility--headings-and-titles"></a>
+
+#### Headings and titles
+
+Use descriptive headings and titles because they help a reader navigate their browser and the
+page. It's easier to jump between pages and sections of a page if the headings and titles are
+unique.
+
+- Use a heading hierarchy.
+- Don't skip levels of the heading hierarchy. For example, put an `h3` element
+  only after an `h2` element.
+- To change the visual formatting of a heading, use CSS rather than using a heading level that
+  doesn't fit the hierarchy.
+- Don't have empty headings or headings with no associated content.
+- Tag headings using heading elements. In HTML: `h1`,
+  `h2`, and so on. In Markdown: `#`, `##`, and so on.
+- Use a level-1 heading for the page title or main content heading.
+
+For more information and examples, see [Headings and titles](#headings).
+
+<a id="accessibility--links"></a>
+
+#### Links
+
+- Use [meaningful link text](#cross-references--descriptive-link-text).
+  Links should make sense when read out of context.
+- Don't use *click here* or *read this document*. Some people who use screen readers
+  jump from link to link to scan a page and need to understand what a link contains.
+- Use *see* to refer to links and cross-references. For more information, see
+  [see](#word-list--see).
+- When a link does anything that the reader might not expect, such as downloading a file,
+  opening in a new tab, or jumping to another section on the same page, explain that behavior when
+  you link. For more information, see
+  [Explain unexpected link behavior](#cross-references--explain-behavior).
+- When possible, avoid adjacent links. Instead, put a character in between to separate them.
+
+<a id="accessibility--lists"></a>
+
+#### Lists
+
+- In a
+  [procedure](#procedures),
+  make each instruction a
+  [list item](#lists).
+- Use lists to make it easier for the reader to follow the steps.
+
+<a id="accessibility--images"></a>
+
+#### Images
+
+- For every image, provide an alt attribute. For alt attributes that contain
+  [alt text](#images--alt-text), use alt text that adequately summarizes the
+  intent of each image. If the image is purely decorative, use empty alt text.
+- Don't present new information in images. Always provide an equivalent text explanation with
+  the image.
+- Don't repeat images unless absolutely necessary.
+- Don't use images of text, code samples, or terminal output. Use actual text.
+- Use SVG instead of PNG if available. SVGs stay sharp when you zoom in on the image.
+
+For more information, see
+[Text associated with images](#images--text-associated-with-images).
+
+<a id="accessibility--videosu002c-recordingsu002c-and-gifs"></a>
+
+#### Videos, recordings, and GIFs
+
+- Provide captions, transcripts, or descriptions of audio and video content. For example, you
+  can use the
+  [autocaption feature](https://support.google.com/youtube/answer/6373554)
+  in YouTube.
+- Ensure that captions can be translated into major languages.
+- Don't use flickering or flashing elements. They can cause anything from motion sickness
+  to a seizure.
+
+<a id="accessibility--buttons-and-icons"></a>
+
+#### Buttons and icons
+
+- For form-submission buttons, use the native HTML `button` element.
+- An icon is a symbol or image that represents an object or a function. For information
+  about using icons, see the [Buttons and icons](#ui-elements--buttons) section
+  of the "UI elements and interaction" page.
+
+<a id="accessibility--ui-navigation"></a>
+
+#### UI navigation
+
+When you use angle brackets (`>`) to document menu paths, add an
+[`aria-label` attribute](https://www.w3.org/TR/WCAG20-TECHS/ARIA14.html)
+to help screen readers interpret the brackets as "and then" instead of as
+"greater than" or "keyboard arrow right". For more information and examples, see
+[Menu bar](#ui-elements--term-menus).
+
+<a id="accessibility--tables"></a>
+
+#### Tables
+
+- Introduce tables in the text preceding the table because not all screen readers preannounce
+  tables.
+- Use table headings for the first column and the first row only. Use the
+  [`th` element](https://www.w3.org/TR/html4/struct/tables.html#edef-TH).
+- If your tables include both row and column headings, then mark heading cells with the
+  [`scope`
+  attribute](https://www.w3.org/WAI/tutorials/tables/two-headers/).
+- If your tables have more than one row containing column headings, then use the
+  [`headers`
+  attribute](https://www.w3.org/WAI/tutorials/tables/multi-level/) and make sure that the headings have unique IDs.
+- Avoid when possible tables in the middle of a numbered procedure.
+- Don't merge cells. Don't use `colspan` or `rowspan` attributes.
+- Don't use tables unless it's the best method to present your information. Tables are
+  challenging for screen readers. For more information, see
+  [List or table](#tables--list-or-table).
+- Don't present new information in tables through images or symbols alone; always provide a
+  descriptive `alt` attribute for the image or symbol. For more information, see
+  [Alt text](#images--alt-text).
+
+For more information, see [Tables](#tables).
+
+<a id="accessibility--interactive-elements"></a>
+
+#### Interactive elements
+
+Introduce an interactive element (such as a button that expands and collapses) in the text
+preceding the element.
+
+Recommended if practical: To see a list of
+requirements, expand the **Requirements** section.
+
+Recommended: To see a list of requirements,
+click the expander arrow.
+
+<a id="accessibility--forms"></a>
+
+#### Forms
+
+- Label every input field by using a `label` element.
+- Place labels outside of fields.
+- When you're creating an error message for form validation, clearly state
+  what went wrong and how to fix it—for example: "Name is a required field."
+
+<a id="accessibility--custom-css-and-javascript"></a>
+
+#### Custom CSS and JavaScript
+
+Try to use your site's standard styles and standard JavaScript code as much
+as possible. However, if you do use custom styles or code, then follow these guidelines:
+
+- Pick colors that respect
+  [accessible color contrast
+  ratios](https://webaim.org/resources/contrastchecker/) (4.5:1 for text).
+- Don't use `visibility:hidden` or `display:none`. Both
+  styles hide information from screen readers.
+- Avoid when possible using mouseover events. But if you do use them, then add alternate
+  focus and blur events for keyboard users.
+- Ensure that any ordering and positioning defined in styles reflects the
+  DOM and the reading order (such as left to right and top to bottom) of your page.
+
+<a id="accessibility--document-rendering"></a>
+
+#### Document rendering
+
+Make sure that your document conveys all the information that you intended when you
+view it in the following contexts:
+
+- Without sound
+- Using only sound
+- Without images, including animation
+- [Without color](https://colororacle.org/)
+- Using a keyboard
+- With screen magnification
+- Without punctuation
+
+Don't use color, size, location, or other visual cues as the primary way
+of communicating information.
+
+- If you're using color, an icon, or outline thickness to convey state,
+  then also provide a secondary cue, such as a change in the text label.
+- Refer to buttons and other elements by their label. For visual elements
+  that have no text, don't try to describe the element. Instead, use the element's
+  `aria-label`
+  attribute if possible.
+  For example:
+
+  Recommended: Click **Save**.
+
+  Recommended: Click **Notifications**.
+
+  Not recommended: Click the bell icon.
+- Don't use directional language to orient the reader, such as *above*, *below*,
+  or *right-hand side*. This type of language doesn't work well for accessibility or for
+  localization reasons. For example, what's on the right side for left-to-right languages
+  appears on the left side for right-to-left languages.
+
+  Don't use directional language to refer to a position in a document. For example, the text
+  isn't *below* if it's being read by a screen reader. Instead, use *earlier*,
+  *preceding*, or *following*.
+
+  Recommended:
+  In the preceding diagram, clients run jobs on multi-team or single-team clusters.
+
+  Not recommended: In the diagram above,
+  clients run jobs on multi-team or single-team clusters.
+
+  If a [UI element](#ui-elements) is hard to find,
+  [provide a screenshot](#images).
+
+  Recommended:
+  Click **Menu**.
+
+  Not recommended: In the left-side
+  panel, click the button with three lines.
+
+<a id="accessibility--more-resources"></a>
+
+#### More resources
+
+- [Google's main
+  accessibility page](https://www.google.com/accessibility/)
+- [Web Content Accessibility
+  Guidelines (WCAG) 2.0](https://www.w3.org/WAI/WCAG20/glance/)
+- [Web Accessibility Initiative
+  (WAI)](https://www.w3.org/WAI/)
+- [Using ARIA](https://www.w3.org/TR/using-aria/)
+- [Web Accessibility
+  Tutorials](https://www.w3.org/WAI/tutorials/)
+
+---
+
+<a id="excessive-claims"></a>
+
+### Excessive claims
+
+*Source: <https://developers.google.com/style/excessive-claims>*
+
+In documentation, don't make excessive claims. An *excessive claim* is an assertion
+in the documentation that does any of the following:
+
+- Makes a statement about performance or cost that isn't easily verifiable with data
+  that's available to the reader.
+- Makes a statement about security that would be invalidated by a security incident.
+- Makes a statement that might be interpreted as subjective or even disparaging,
+  especially about third-party products.
+
+When you're assessing whether some text makes an excessive claim, take into account
+not just what's true today about a product's performance, cost, security, or
+functionality, but what might be true in the future.
+
+Consider the following guidelines:
+
+- When you describe products, avoid superlatives like *best*, *simplest*,
+  *fastest*, *never*, and *always*. Similarly, be
+  careful about words like *ensure* and *guarantee* and use them only when
+  something can truly be ensured or guaranteed.
+- If you make specific performance claims—how fast a product is, how much storage
+  it requires, and so on—make sure that you reference the source of your information.
+- If documentation claims that a product is secure, the documentation
+  is invalid (and not credible) if someone succeeds in compromising the product.
+  It's safer to suggest that a feature "helps with security" or "is designed for
+  security" because those statements are true even if a security incident occurs.
+- A statement that you make about a competitive product might be untrue if you
+  misinterpret how the product works, or later if the other company comes out with
+  a new release.
+
+The safest approach is always to write factually and objectively, limiting what you say to
+verifiable information that will be true over the lifespan of your documentation.
+
+Recommended: Our product
+distributes datasets and computation in memory across a cluster, and
+therefore it can be faster for this scenario than ExampleCorporation's product. For
+more information, see [Performance comparison](https://www.google.com/).
+
+Not recommended: Our product is
+faster than ExampleCorp's product.
+
+Recommended: Using our security product
+is part of an overall strategy that helps prevent account takeovers from phishing attacks.
+
+Not recommended: Our security product
+prevents account takeovers from phishing attacks.
+
+---
+
+<a id="future"></a>
+
+### Future features
+
+*Source: <https://developers.google.com/style/future>*
+
+Avoid documenting future features or products, even in innocuous
+ways. Don't pre-announce anything in documentation unless it has been approved by your legal counsel.
+
+See also [Present tense](#tense) and
+[Timeless documentation](#timeless-documentation).
+
+---
+
+<a id="translation"></a>
+
+### Global audience
+
+*Source: <https://developers.google.com/style/translation>*
+
+We write our developer documentation in US English, but some of it is
+translated into languages other than English or is read by developers for whom
+English is not their primary language.
+
+Write with localization, translation, and
+internationalization in mind. The following list defines these terms:
+
+- *Localization:* Adapting a product and its associated documentation for a specific country.
+  This process involves more than translation—for example, using local currencies or units of
+  measurement.
+- *Translation:* Translating one language to another language. This process might involve
+  localization, but the two terms aren't synonymous with one another.
+- *Internationalization:* Designing a product and its associated documentation to minimize
+  the localization effort—for example, placing all UI strings in a separate file to simplify
+  translation.
+
+For more information, see
+[Language localization](https://wikipedia.org/wiki/Language_localisation).
+
+For other writing best practices, see the following resources:
+
+- [Write accessible documentation](#accessibility)
+- [Write inclusive documentation](#inclusive-documentation)
+- [Voice and tone](#tone)
+
+<a id="translation--use-clearu002c-conciseu002c-and-unambiguous-language"></a>
+
+#### Use clear, concise, and unambiguous language
+
+Consider global audiences and translation and write in a way that's clear, concise, and
+unambiguous.
+
+<a id="translation--use-simpler-words-and-shorter-sentences"></a>
+
+##### Use simpler words and shorter sentences
+
+- Use a simple word. For example, don't use words like *commence* when you mean *start*
+  or *begin*. Don't use *consequently* when you mean *so*. Don't use words like
+  *utilize* or *leverage* when you mean *use*. (It's fine to use these words when
+  you're conveying a special sense—for example, *Cloud Spanner utilizes up to 100% of the available
+  CPU resources.*)
+- Use a single word when it conveys the same idea as a phrase. For example, don't
+  use a phrase like *a number of* when you can use *some* or *many*.
+- Write shorter sentences. The shorter the sentence, the easier it is to translate. English sentences can be
+  shorter in length than some languages, so an English sentence of average length might result in a
+  long sentence when translated. Longer sentences can impair understanding, cause rendering issues
+  on the page or product interface, lengthen translation time, and increase translation and
+  review costs.
+
+<a id="translation--avoid-phrasal-verbs"></a>
+
+##### Avoid phrasal verbs
+
+- Avoid phrasal verbs when possible. A phrasal verb combines multiple words to form a single
+  verb phrase. These verbs are also known as compound verbs. Try to substitute a simpler verb first.
+  There might not be a better verb; for example, a few exceptions to this rule include *set up*,
+  *log in*, and *sign in*.
+
+  Recommended: This document uses the following
+  terms:
+
+  Not recommended: This document makes use of
+  the following terms:
+
+<a id="translation--use-modifiers-appropriately"></a>
+
+##### Use modifiers appropriately
+
+- Don't use too many modifiers. In particular, don't use more than two nouns as modifiers of
+  another noun.
+
+  Recommended: A cloud-native DevSecOps
+  pipeline in a hybrid environment
+
+  Not recommended: A hybrid cloud-native
+  DevSecOps pipeline
+- Don't misplace modifiers. For example, place a word like *only* immediately before the
+  word or phrase that it relates to. If the meaning is still ambiguous, try rephrasing the sentence.
+
+  Recommended: Request only one token.
+
+  Recommended: Request no more than one token.
+
+  Not recommended: Only request one token.
+
+<a id="translation--use-active-voice-and-present-tense"></a>
+
+##### Use active voice and present tense
+
+- Use [present tense](#tense) and avoid complex or uncommon verb forms.
+- Use active voice. The subject of the sentence is the person or thing performing the action.
+  With passive voice, it's often hard for readers to figure out who's supposed to do something.
+  For more information, see [Active voice](#voice).
+
+<a id="translation--use-words-in-their-primary-sense"></a>
+
+##### Use words in their primary sense
+
+- Don't use the same word to mean different things. In particular, avoid using the same word as
+  both a noun and a verb in close proximity. For examples of words that have multiple meanings, see the word
+  list entries for [once](#word-list--once), [while](#word-list--while), [as](#word-list--as), and [since](#word-list--since).
+- Avoid directional language (for example, *above* or *below*) in procedural
+  documentation. For more information, see
+  [UI elements and interaction](#ui-elements--buttons).
+
+<a id="translation--use-helper-words-and-optional-words"></a>
+
+##### Use helper words and optional words
+
+- Use qualifying nouns for technical keywords. For example, when referring to a file called
+  `example.yaml`, call it the *`example.yaml` file* and not
+  *`example.yaml`* by itself. For more information, see
+  [Grammatical treatment of code elements](#code-in-text--keywords).
+- Repeat a word if the redundancy improves comprehension.
+
+  | Recommended | Not recommended |
+  | --- | --- |
+  | If the VM has started and if you're able to connect... | If the VM has started and you're able to connect... |
+  | The resource hierarchy design creates both IAM segmentation and network segmentation by default. | The resource hierarchy design creates both IAM and network segmentation by default. |
+  | An egress rule whose action is `allow`, whose destination is `0.0.0.0/0`, and whose priority is the lowest possible (`65535`). | An egress rule whose action is `allow`, destination is `0.0.0.0/0`, and priority is the lowest possible (`65535`). |
+- Use helper words. Helper words such as *then*, *that*, and *of*
+  are frequently left out of conversational English. Use these words to avoid ambiguity.
+
+  | Recommended | Not recommended |
+  | --- | --- |
+  | If the attribute key is not found, then the default value is returned. | If the attribute key is not found, the default value is returned. |
+  | This document is intended for data engineers and assumes that you have the following knowledge: | This document is intended for data engineers and assumes you have the following knowledge: |
+  | Identify all of the datasets. | Identify all the datasets. |
+  | Start the profiler, and then run the app. | Start the profiler, then run the app. |
+
+  See also [Optional pronouns](#pronouns--optional-pronouns).
+- Don't omit relative pronouns. To provide clarity and to avoid ambiguity, use relative
+  pronouns such as *that* and *which*. For more information, see
+  [Relative pronouns](#pronouns--relative-pronouns).
+
+  Recommended: You can programmatically update
+  the rules that you previously defined.
+
+  Not recommended: You can programmatically
+  update the rules you previously defined.
+
+<a id="translation--clarify-abbreviations-and-pronouns"></a>
+
+##### Clarify abbreviations and pronouns
+
+- Define abbreviations. Abbreviations can be confusing out of context, and they don't translate
+  well. Spell things out whenever possible, at least the first time that you use
+  a given term. For more information, see [Abbreviations](#abbreviations).
+- Clarify antecedents. Using pronouns can get tricky when translators are working with small,
+  unconnected strings of text. Help them out by making things as clear as
+  possible. For example, if a pronoun is ambiguous, then replace it with the
+  appropriate noun.
+
+  Recommended: If you use the term
+  *green beer* in an ad, then make sure that the ad is targeted.
+
+  Not recommended: If you use the term
+  *green beer* in an ad, then make sure that it's targeted.
+
+<a id="translation--use-apostrophes-appropriately"></a>
+
+##### Use apostrophes appropriately
+
+Be careful with how you use plural and possessive forms. In general, don't form a plural with
+*'s*, don't use the plural or possessive form with trademarks of company, product, and feature
+names, and don't use uncommon contractions. For more information, see [Possessives](#possessives), [Pluralization](#pluralization), and [Contractions](#contractions).
+
+<a id="translation--address-users-and-their-needs-directly"></a>
+
+#### Address users and their needs directly
+
+Address the user and their needs directly and avoid providing unnecessary information.
+
+- Address the reader directly. Use *you*, instead of *the user* or *they*, unless
+  you're referring to someone who uses the software that the reader is developing. For more
+  information, see [Second person and first person](#person).
+- Provide context. Don't assume that the reader already knows what you're talking about.
+- Avoid negative constructions when possible. Consider whether it's necessary to tell the reader
+  what they can't do instead of what they can.
+
+<a id="translation--be-consistent"></a>
+
+#### Be consistent
+
+Use standard sentence structures, consistent terminology, and appropriate punctuation to avoid
+creating barriers to understanding, ambiguity, and mistranslations.
+
+<a id="translation--use-consistent-terminology"></a>
+
+##### Use consistent terminology
+
+If you use a particular term for a concept in one place, then use that exact same term
+elsewhere, including the same capitalization. If you use different names for the same thing,
+translators might think you're referring to different concepts, and thus might use different
+translations. Inconsistency in terminology and phrasing can increase translation costs,
+particularly when translation memory and machine translations are used as first steps in
+translation.
+
+<a id="translation--use-standard-sentence-structures-and-formatting"></a>
+
+##### Use standard sentence structures and formatting
+
+- Use standardized phrases for frequently used sentences, introductory phrases, and other common
+  tasks. For examples, read about [introducing links](#cross-references--link-introductions),
+  [introducing output](#placeholders--placeholders-in-output), and
+  [introducing code samples](https://developers.google.com/style/code-samples#introductions).
+- Use standard English word order. Sentences follow the *subject + verb + object* order.
+- Try to keep the main subject and verb as close to the beginning of the sentence as possible.
+- Use the conditional clause first. If you want to tell the audience to do something in a
+  particular circumstance, mention the circumstance before you provide the instruction. For more
+  information, see [Sentence structure](#sentence-structure).
+- Make list items consistent. Make list items parallel in structure. Be consistent in your
+  capitalization and punctuation. For more information, see [Lists](#lists).
+
+<a id="translation--use-consistent-text-formatting"></a>
+
+##### Use consistent text formatting
+
+- Use consistent typographic formats. Use bold and italics consistently. Don't switch from
+  using italics for emphasis to underlining. For more information, see
+  [Text-formatting summary](#text-formatting).
+- Use consistent capitalization. For more information, see
+  [Capitalization](#capitalization).
+
+<a id="translation--be-inclusive"></a>
+
+#### Be inclusive
+
+You're not writing for your culture. Write with inclusivity in mind. For more information, see [Writing inclusive documentation](#inclusive-documentation).
+
+- Write [dates and times](#dates-times) in unambiguous and clear ways.
+- Don't be too
+  culturally specific. In particular, don't refer to specific holidays, cultural practices, or sports
+  unless you're certain they're known worldwide.
+- Use a diverse set of example names. If you
+  need to use people's names (for example, as email addresses), use a diverse set of names. For more
+  information, see [Example domains and names](#examples).
+- Avoid
+  colloquialisms, idioms, or slang. Phrases like *ballpark figure*, *back burner*, or
+  *hang in there* can be confusing and difficult to translate.
+- Avoid humor. Most humor
+  is difficult to translate, and much humor is culturally specific.
+- Avoid geographically
+  specific references, like the seasons. Remember that August isn't summer in the southern hemisphere.
+  For more information, see [Expressing divisions of the
+  year](#dates-times--divisions-year).
+
+<a id="translation--consider-accessibility-for-images"></a>
+
+#### Consider accessibility for images
+
+Use screenshots and text in figures sparingly. Images don't get translated. Any new information
+should be conveyed through text and not introduced in a figure or image. For more information, see
+[Figures and other images](#images).
+
+---
+
+<a id="inclusive-documentation"></a>
+
+### Inclusive language
+
+*Source: <https://developers.google.com/style/inclusive-documentation>*
+
+> [!NOTE]
+> **Note**: This document includes references to
+> potentially disrespectful or offensive terms. These terms are listed here to
+> provide usage guidance and alternative terms.
+
+When you write developer documentation with inclusivity and diversity in mind,
+you help ensure that the content is more precise and clear for all readers.
+Avoid any kind of idiomatic or figurative language that can be misinterpreted or
+distracting.
+
+This page is not an exhaustive reference, but provides some general guidelines
+and examples that illustrate some best practices for writing inclusive
+documentation.
+
+For other writing best practices, see the following resources:
+
+- [Write for a global audience](#translation)
+- [Write accessible documentation](#accessibility)
+- [Voice and tone](#tone)
+
+<a id="inclusive-documentation--gendered-language"></a>
+
+#### Avoid unnecessarily gendered language
+
+Be mindful of the
+[pronouns](#pronouns--gender-neutral-pronouns)
+that are used in narrative
+examples, and be aware of other possible sources of gendered language.
+
+| Recommended | Not recommended |
+| --- | --- |
+| Equipment installation takes around 16 person-hours to complete. | Equipment installation takes around 16 man-hours to complete. |
+| Build AI that benefits humanity. | Build AI that benefits mankind. |
+
+<a id="inclusive-documentation--figurative-language"></a>
+
+#### Avoid figurative language
+
+Use simple language and terminology that's precise and clear for all of your
+audiences:
+
+- Avoid idiomatic or figurative language that can be misunderstood,
+  distracting, or difficult for translation.
+- Avoid [jargon](#jargon).
+- Use terms that are established industry standards and widely understood by
+  the target audience.
+
+When you try to achieve a [friendly and conversational
+tone](#tone), you might mistakenly use figurative language. Figurative language can
+come in the form of figures of speech and other turns of phrase. Be attentive to
+your word choice, especially when you aim for an informal tone.
+
+Don't use metaphors, and don't use a term in a metaphorical sense
+([use words in their primary sense](#translation--use-words-in-their-primary-sense)).
+For example, avoid using the metaphor of *pets versus cattle* when you
+compare on-premises or stateful systems with stateless cloud systems.
+
+For guidance about specific terms, see the
+[Word list](#word-list).
+
+<a id="inclusive-documentation--ableist-language"></a>
+
+##### Avoid ableist language
+
+Ableist language includes
+words or phrases such as *crazy*, *insane*, *blind to* or
+*blind eye to*, *cripple*, *dumb*, and others. Choose a more
+accurate or alternative word, depending on the context.
+
+| Recommended | Not recommended |
+| --- | --- |
+| Before launch, give everything a final check for completeness and clarity. | Before launch, give everything a final sanity-check. |
+| There are some baffling outliers in the data. | There are some crazy outliers in the data. |
+| It slows down the service, causing a poor user experience until the queue clears. | It cripples the service, causing a poor user experience until the queue clears. |
+| Replace the placeholder in this example with the appropriate value. | Replace the [dummy variable](#word-list--dummy-variable) in this example with the appropriate value. |
+
+<a id="inclusive-documentation--graphic-language"></a>
+
+##### <a id="inclusive-documentation--violent-language"></a>Avoid graphic or metaphorical language
+
+Avoid unnecessarily graphic or metaphorical language, when you can use a more
+precise term.
+
+For example, instead of
+*[STONITH](#word-list--stonith)*, use specific terms to
+describe the process that's used to stop an errant node. If you need to mention a term
+such as *STONITH*, you can mention it once when you first explain the
+relevant feature, and phrase it in a way that de-emphasizes the term.
+
+Recommended:
+This approach might require you to fence failed nodes.
+
+Sometimes okay:
+This approach might require you to fence failed nodes (sometimes referred to
+as *STONITH*).
+
+Always use the most precise and well-understood terms for your context.
+In some contexts, an industry-established term has a specific technical meaning
+that doesn't have an accurate synonym or alternative. For examples, see the word
+list entries for [terminate](#word-list--terminate) and
+[execute](#word-list--execute).
+
+| Recommended | Not recommended |
+| --- | --- |
+| If the connection doesn't respond, check for errors. | If the connection hangs, check for errors. |
+| Point to **File**, and then click **New**. | Hover over **File**, and hit **New**. |
+
+For guidance about specific terms, see the
+[Word list](#word-list).
+
+<a id="inclusive-documentation--diverse-examples"></a>
+
+#### Write diverse and inclusive examples
+
+Write documentation for a
+[global audience](#translation--be-inclusive).
+Use diverse names, genders, ages, and locations in examples. Keep the following
+advice in mind:
+
+- Follow our [gender-neutral
+  pronoun](#pronouns--gender-neutral-pronouns) guidance.
+- Avoid being too culturally specific to the US. Be mindful when referring
+  to specific holidays (see also the word list entry for [*the holidays*](#word-list--holiday)), cultural practices,
+  sports, and figures of speech.
+- In examples,
+  [choose a diverse set of names](#examples--example-person-names)
+  to help reflect our global audience. For guidelines about fictional people,
+  see
+  [Further notes about example people](#examples--further-notes-about-example-people).
+- When writing about older adults, avoid terms and figures of speech such
+  as *the elderly*, *the aged*, *seniors*,
+  *senior citizens*, or *80 years young*. Instead, use terms such as
+  *older adults* or *aging population*, or mention the person's
+  relative age or relationship to the other people in your example when those
+  details are relevant.
+
+<a id="inclusive-documentation--features-and-users"></a>
+
+#### Write about features and users in inclusive ways
+
+Avoid referring to people in divisive ways. For example, instead of referring
+to people as *native speakers* or *non-native speakers* of English, consider
+whether your document needs to discuss this at all, and revise it
+to discuss the feature in terms that are relevant to anyone regardless of what
+languages they know.
+
+Avoid using socially charged terms for technical concepts where possible. For
+example, avoid terms such as [blacklist](#word-list--blacklist) and
+[native](#word-list--native) feature, and don't use terms like
+[first-class
+citizen](https://wikipedia.org/wiki/First-class_citizen), even though such terms might still be widely used.
+
+<a id="inclusive-documentation--replace-or-write-around-non-inclusive-terms"></a>
+
+##### Replace or write around non-inclusive terms
+
+This section contains guidance about how to replace or write around a non-inclusive term. If a
+term is well established in the industry and replacing it could cause confusion, see
+[Replace established terms](#inclusive-documentation--replace). If a term occurs in code samples or keywords, see
+[Write around non-inclusive code terms](#inclusive-documentation--write-around). For information about avoiding
+non-inclusive jargon, see [Jargon](#jargon).
+
+<a id="inclusive-documentation--replace"></a>
+
+###### Replace established terms
+
+Many non-inclusive terms are in wide use in the industry, such as *whitelist*. If replacing
+an established term could cause confusion for readers, you can directly refer to the non-inclusive
+term on the first use, and put it in parentheses. Then use the inclusive, replacement term
+throughout the rest of the document.
+
+Recommended: To make sure that administrators
+get the notification, add them to an allowlist (sometimes called a *whitelist*). Anyone who
+isn't on the allowlist is blocked ...
+
+Recommended: In this model, a Jenkins
+controller (master) handles HTTP requests. The Jenkins controller is designed to ...
+
+Recommended: In cloud architecture, servers
+are treated as commodities (sometimes described by using the metaphor *cattle, not pets*).
+
+In many cases, instead of directly replacing a word, you can rewrite to improve the clarity of a
+sentence. For example, instead of replacing the verb *whitelist* with *allowlist*, try
+rewriting the sentence.
+
+Recommended: You can allow requests from a
+range of IP addresses by entering a CIDR block instead of a single address in the field.
+
+Not recommended: You can allowlist a range of
+IP addresses by entering a CIDR block instead of a single address in the field.
+
+<a id="inclusive-documentation--write-around"></a>
+
+###### Write around non-inclusive code terms
+
+In some cases, non-inclusive terms are embedded in code (or similar) as names or keywords, and
+you can't simply ignore those terms and use different terminology. What you can do, however, is
+*minimize* your use of the term (hence avoid propagating it as a term of art), while still
+providing clear documentation to your readers. Don't use a non-inclusive name or keyword unless it's
+in code font.
+
+Following are scenarios for writing around non-inclusive terms that occur in code and keywords.
+
+One scenario is if you're documenting an existing system in which an entity is already named
+by using a non-inclusive term. For example, there might be a configuration file that includes the
+following cluster name:
+
+```
+apiVersion: v1
+kind: Config
+preferences: {}
+
+clusters:
+- cluster:
+  name: master
+- cluster:
+  name: replica-1
+```
+
+Another scenario is if your documentation includes a non-inclusive term that's an established
+keyword, such as the keyword `SLAVE` in dialects of SQL:
+
+```
+START SLAVE UNTIL SQL_AFTER_MTS_GAPS;
+```
+
+The first time that you refer to a code item that uses a non-inclusive term, you can directly
+refer to that term, but format it in code font, and put it in parentheses if possible.
+
+Recommended: The configuration file helps you
+create a parent node (which is named `master` in the file).
+
+Recommended: Start the replica by using the
+`START SLAVE` statement.
+
+In subsequent mentions, use the preferred term (*parent node*, *replica*). If it's
+necessary to refer to the entity name or keyword, continue doing so only with code formatting.
+
+<a id="inclusive-documentation--about-disability-and-accessibility"></a>
+
+#### Avoid bias and harm when discussing disability and accessibility
+
+Many developers create products with accessibility and disability in mind.
+When documenting these features, and when writing about people with
+disabilities or about accessibility, work to eliminate unintentional bias and
+harm. Take the time to educate yourself about the ways that the communities that you're
+writing about prefer to be identified and described before writing about them in
+your documentation.
+
+Some general guidelines in this area include the following:
+
+- Don't describe people without disabilities as *normal* or *healthy*. This
+  contributes to othering and alienation of people with disabilities by implying that
+  they are abnormal or sick. Instead, use terms such as *nondisabled person*,
+  *sighted person*, *hearing person*,
+  *person without disabilities*, or *neurotypical person*.
+- Research the ways that the people in the communities that you're writing about
+  prefer to be identified and use the terms that they prefer. In many cases, avoid
+  terms that remove personhood or that define people by their disability. For
+  example, avoid terms such as *the disabled* or *a quadriplegic*.
+  Instead, use terms such as *people with disabilities* or *a quadriplegic person*.
+
+  However, many members of some communities prefer *identity-first language*—for
+  example, that preference is common in autistic, blind, and Deaf communities. Capitalization of
+  identities also can vary (for some perspectives, visit
+  [Identity-First Language](https://autisticadvocacy.org/about-asan/identity-first-language/)
+  and
+  [Self-Identification
+  in the Deaf Community](https://www.verywellhealth.com/deaf-culture-big-d-small-d-1046233)). Whenever possible, research and choose terms
+  that respect the ways that people in the communities identify.
+- Use *see* to refer to links and cross-references. For more information, see
+  [see](#word-list--see).
+- Avoid terms that reflect or project feelings and judgments about a person's disability,
+  such as *victim of*, *suffering from*, or *wheelchair-bound*. Instead, use neutral
+  terms such as *experiencing*, *living with*, or *uses a wheelchair*.
+- Avoid euphemisms or patronizing terms such as *physically challenged*, *special*,
+  *differently abled*, or *handi-capable*.
+
+---
+
+<a id="jargon"></a>
+
+### Jargon
+
+*Source: <https://developers.google.com/style/jargon>*
+
+Jargon is the specialized and often figurative terminology of a specific group to represent a
+larger concept—for example, *camel case*, *swim lane*,
+*break-glass procedure*, or *out-of-the-box*. Jargon can also include
+vaguely defined or overloaded terms like *solution*, *support*, or
+*workload*.
+
+Typically, the meaning of jargon isn't understood except by the specific group. For this reason,
+jargon can hamper our efforts to publish content that's clear, that reaches a
+[global audience](#translation)
+in multiple languages, that serves readers at various levels of product knowledge, and that's
+inclusive of different groups and cultures. For more information about writing with
+inclusivity and diversity in mind, see
+[Write inclusive documentation](#inclusive-documentation).
+
+However, some jargon is widely understood and accepted by our industry or by the intended
+audience of a document. It can be valuable to include jargon in a document when you know that
+readers search for those terms. If you're going to use jargon, consider the following questions:
+
+- **Can you write around the term?** If you don't need the term for search engine
+  optimization (SEO), try writing around it. For example, instead of writing *Hold a
+  post-mortem*, write *When the project is finished, review what processes worked or didn't
+  work*. Instead of writing *Create a back-of-the-envelope design*, write *Use an informal
+  design process*.
+- **Can you replace the term with a different, more specific term?** For example, the
+  [word list](#word-list)
+  for this style guide offers several replacement terms: *affected area* or *spatial
+  impact* (for *blast radius*), *import* or *load* (for *ingest*), and
+  *ready-made* or *pre-built* (for *off-the-shelf*). When a term on the word list is
+  marked as "Don't use" (some jargon can be considered offensive, violent, or not inclusive),
+  replace that term or write around it.
+- **Are you using the term only once in your document?** If so, describe the term in plain
+  language and refer to it in parentheses, or link to a trusted definition.
+
+  Recommended: You then move the task to an
+  earlier part of the process (also known as *shifting left*).
+
+  Recommended: A
+  [split-brain](https://en.wikipedia.org/wiki/Split-brain_(computing))
+  situation can develop.
+- **Are you using the term throughout your document?** If so, briefly describe the term in
+  parentheses on first reference, or link to a trusted definition.
+
+  Recommended: The application is in the
+  same state as a *cold standby* (a backup or redundant system that's identical to a primary
+  system).
+
+  Recommended: A better approach is to use
+  a pattern called a
+  [*dead letter queue*](https://en.wikipedia.org/wiki/Dead_letter_queue).
+- **Is the term used in a command or code sample?** If so, use the words only in direct reference to the code items
+  ([formatted as code](#code-in-text)), and make it clear
+  what you're referring to.
+
+  Recommended: Add a user to the
+  allowlist (`whitelist`) by entering the following:
+  `whitelist adduser EMAIL_ADDRESS`.
+
+  Not recommended: Add a user to the
+  whitelist by entering the following: `whitelist adduser
+  EMAIL_ADDRESS`.
+
+---
+
+<a id="prescriptive-documentation"></a>
+
+### Prescriptive documentation
+
+*Source: <https://developers.google.com/style/prescriptive-documentation>*
+
+Write prescriptive documentation.
+
+*Prescriptive* (or *opinionated*) documentation recommends a way to achieve tasks
+and accomplish goals. It tells the reader what to do instead of giving them a list of options to
+choose from. When a goal or task is complex and involves multiple approaches or products,
+prescriptive documentation recommends a path.
+
+Prescriptive writing affects several aspects of documentation:
+
+- **The purpose and structure of a document**. Prescriptive documentation states a clear,
+  specific purpose. Headings and content are written with that purpose in mind.
+- **Example scenarios and procedures**. Scenarios and procedures reflect the use cases that
+  are most likely relevant to the readers.
+- **Sample commands**. Prescriptive documentation provides commands and arguments that
+  accomplish the task for the most common use case. For more information about documenting
+  command-line options, see
+  [Optional arguments in click-to-copy commands](#code-syntax--click-to-copy-commands).
+
+For instance, best practice documents are typically prescriptive documents. For an example, see
+[Operations best practices](https://cloud.google.com/architecture/security-foundations/operation-best-practices).
+
+<a id="prescriptive-documentation--word-choice"></a>
+
+#### Word choice for recommendations and requirements
+
+To indicate required or optional user actions or the outcomes of a process, select an appropriate
+auxiliary verb—for example, *must*, *can*, or *might*. Generally avoid the word
+*should*. The word can create ambiguity and uncertainty for readers and is thus problematic for
+prescriptive documentation. For example, if you're telling the reader what to do, *should*
+implies that the action is recommended but optional, which can leave the reader unsure about what to
+do.
+
+To clarify what you mean, determine if an action is *required* versus *optional*, an
+outcome is *expected* versus *possible*, or a state is *actual* versus
+*recommended*.
+
+- **If an action is required**: use *must*, or rephrase
+  the sentence so that it's a clear imperative instruction such as
+  "Do the following before you continue."
+- **If an action is recommended**: use *We recommend ...* or
+  *Google recommends ...*. You can use *should* if a
+  recommended action is generally recognized. For example, "You should
+  use a strong password ..." or "You should follow the principle of
+  least privilege ...."
+- **If an action is optional**: use *can*. For example,
+  "You can also use approach B to solve the same problem."
+- **If an outcome is expected**: describe the outcome in terms of
+  what is expected. For example: "The process returns 10 items."
+- **If an outcome is possible**: use *might* or *can*.
+  For example, "The process can take about 30 minutes."
+- **If a state is actual**: when you're describing the state of
+  something, such as the value of a variable, avoid writing "The value
+  should be true." Instead, clarify which of the following you mean:
+  - "You must set the value to true."
+  - "The server sets the value to true."
+  - "If the value is false, follow these steps to change it to true."
+
+  For information about clarifying who's performing an action, see
+  [Active voice](#voice).
+
+Recommended: Ensure that the
+Classroom Share Button conforms to our min-max size guidelines and related
+color/button templates.
+
+Recommended: The column of the data
+table that the filter operates on.
+
+Recommended: Whether it's a brand new
+project or an existing one, perform the following steps.
+
+Not recommended: The Classroom Share
+Button should conform to our min-max size guidelines and related color and
+button templates.
+
+Not recommended: The column of the
+data table that the filter should operate on.
+
+Not recommended: Whether it's a brand
+new project or an existing one, here's what you should do.
+
+<a id="prescriptive-documentation--more-resources"></a>
+
+#### More resources
+
+- See also [can](#word-list--can), [could](#word-list--could),
+  [may](#word-list--may), [might](#word-list--might),
+  [must](#word-list--must), and [would](#word-list--would) in the
+  word list.
+
+---
+
+<a id="other-sources"></a>
+
+### Third-party content
+
+*Source: <https://developers.google.com/style/other-sources>*
+
+Don't copy content from another source because it might violate copyright. Instead, paraphrase
+and link to their content.
+
+Content includes the following types: text, images, code, logos, and speech.
+
+Recommended: A
+[recovery point objective (RPO)](https://en.wikipedia.org/wiki/Disaster_recovery#Recovery_Point_Objective),
+which is the maximum acceptable length of time during which data might be lost from your app due to
+a major incident.
+
+Not recommended: Recovery Point Objective (RPO): "RPO is the
+maximum targeted period in which data (transactions) might be lost from an IT service due to a major
+incident" (<https://en.wikipedia.org/wiki/Disaster_recovery#Recovery_Point_Objective>).
+
+<a id="other-sources--third-party"></a>
+
+#### Avoid third-party content
+
+Unless you are sure that your company owns the assets, avoid copying from these sources:
+
+- Third-party sources: This list includes documentation, websites, books, blogs, videos, images,
+  podcasts, and more.
+- Reference sources: Avoid copying from dictionaries, encyclopedias, and Wikipedia.
+- Open source product documentation: Open source software (OSS) has different license options,
+  which can range from no reuse without attribution to complete freedom to use the material. It's
+  not safe to assume that you can reuse this content freely. When in doubt, don't use their
+  content.
+- GitHub content: Different GitHub users might adopt different licenses for their content. It's
+  not safe to assume that you can reuse this content freely. When in doubt, don't use their
+  content.
+
+---
+
+<a id="timeless-documentation"></a>
+
+### Timeless documentation
+
+*Source: <https://developers.google.com/style/timeless-documentation>*
+
+Timeless documentation is documentation that avoids words and phrases that anchor the
+documentation to a point in time or assume knowledge of prior or future products and features. In
+general, document the current version of a product or feature.
+
+Timeless documentation is especially important for technical documents that might be read a long
+time after they are written. Words like *now*, *new*, and *currently* can render
+such documentation inaccurate, outdated, or unmeaningful. In contrast, timeless documentation
+focuses on how the product works right now—not on how it has changed from previous versions,
+and not how it might change in the future.
+
+| Recommended | Not recommended |
+| --- | --- |
+| These subcommands let you interact with HTTP load balancing. | These new subcommands let you interact with HTTP load balancing. |
+| The following command-line options aren't supported: | The following command-line options aren't currently supported: |
+| The emulator supports the following filters: | The emulator now supports the following filters: |
+
+If you're writing procedural or time-stamped content such as press releases, blog posts, or
+release notes, such time-based words and phrases are okay. For example, *new* is okay in a blog
+post that announces updates to a product: *Dataflow includes several new features.* Or,
+*soon* is okay in procedural content to emphasize a change in state after a user performs a
+step: *The VM goes offline soon after you send the shutdown command.* However, some of these
+words can become outdated or incorrect when used in product documentation to refer to a product's
+features and capabilities, so we recommend against using such words in that context.
+
+Writing timeless product documentation has the following value:
+
+- It reduces the maintenance required to keep documentation up to date.
+- It avoids assuming the reader is familiar with earlier versions of the product.
+
+<a id="timeless-documentation--words-and-phrases-to-avoid"></a>
+
+#### Words and phrases to avoid
+
+The following words and phrases can undermine timelessness in documentation:
+
+- **Words and phrases that make promises or project plans and
+  strategies**. In the context of describing product or feature capabilities, words and phrases such
+  as *at present*, *as of this writing*, or *eventually* can prematurely disclose plans
+  for a product or feature, or they can inappropriately imply that a product or feature might change.
+  In those cases, don't use such words and phrases.
+
+  For more information, see [Documenting future features](#future).
+- **Words and phrases that are implied**. At Google, we assume our documentation is
+  current unless a specific release version is specified. Thus, words and phrases such as
+  *currently* and *as of this writing* are implied by the existence of the documentation
+  itself.
+- **Words and phrases that become outdated soon after publication**. Words such as *soon*
+  and *latest* quickly become irrelevant.
+- **Words and phrases that assume prior knowledge of a product or feature**. If you must use
+  words like *new*, give a reference point such as a date or version release number—for
+  example, *The January 14, 2021 release of BigQuery includes a new resource panel.*
+
+When describing product or feature capabilities in product and reference documentation, avoid
+the following words and phrases:
+
+- as of this writing
+- currently
+- does not yet
+- eventually
+- existing
+- future, in the future
+- latest
+- new, newer
+- now
+- old, older
+- presently, at present
+- soon
+
+---
+
+<a id="tone"></a>
+
+### Voice and tone
+
+*Source: <https://developers.google.com/style/tone>*
+
+In your documents, aim for a voice and tone that's conversational, friendly,
+and respectful without using slang or being overly colloquial or frivolous. Use
+a voice that's casual, natural, and approachable, not pedantic or pushy. Try to
+sound like a knowledgeable friend who understands what the developer wants to do.
+
+Don't try to write exactly the way you speak; you probably speak more
+colloquially and verbosely than you should write, at least for developer
+documentation. But, aim for a conversational tone rather than a formal one.
+
+Don't try to be super-entertaining, but also don't aim for super-dry. Be
+human, let your personality show, and be memorable. But remember that the
+primary purpose of the document is to provide information to someone who's
+looking for it and may be in a hurry.
+
+Consider that readers come from many different cultures and may have varying
+levels of ability reading English. As much as possible, avoid culturally
+specific references. Simple and consistent writing can also make it easier to
+translate documents into other languages. For more information, see
+[Writing for a global audience](#translation).
+
+For other writing best practices, see the following resources:
+
+- [Write accessible documentation](#accessibility)
+- [Write inclusive documentation](#inclusive-documentation)
+
+<a id="tone--avoid"></a>
+
+<a id="tone--some-things-to-avoid-where-possible"></a>
+
+#### Some things to avoid where possible
+
+- Buzzwords or
+  [technical jargon](#jargon).
+- Being too cutesy.
+- [Avoid figurative language](#inclusive-documentation--figurative-language),
+  which includes metaphors and ableist language.
+- Placeholder phrases like *please note* and *at this time.*
+- Choppy or long-winded sentences.
+- Starting all sentences with the same phrase (such as *You can* or *To
+  do*).
+- Current pop-culture references.
+- Exclamation marks. In general, avoid exclamation points. See [Specific guidance on exclamation points](#periods--exclamation-points).
+- Wackiness, zaniness, and goofiness.
+- Phrasing that denigrates or insults any group of people.
+- Phrasing in terms of *let's* do something.
+- Using phrases like *simply*, *It's that simple*, *It's easy*, or *quickly* in a
+  procedure.
+- Internet slang, or other [internet
+  abbreviations](#abbreviations--dont-use) such as *[tl;dr](#word-list--tldr)* or
+  *[ymmv](#word-list--ymmv)*.
+
+<a id="tone--techniques"></a>
+
+<a id="tone--some-techniques-and-approaches-to-consider"></a>
+
+#### Some techniques and approaches to consider
+
+- If you're having trouble expressing something, step back and ask yourself,
+  "What am I trying to say?" Often, the answer you give yourself reveals what you
+  should be saying in the document.
+- If you're uncertain about your phrasing or tone, ask a colleague to take a
+  look.
+- Try reading parts of your document out loud, or at least mouthing the
+  words. Does it sound natural? Not every sentence has to sound natural when
+  spoken; these are written documents. But if you come across a sentence that's
+  awkward or confusing when spoken, consider whether you can make it more
+  conversational.
+- Use transitions between sentences. Phrases like *Though* or *This way* can
+  make paragraphs less stilted. (Then again, sometimes transitions like *However*
+  or *Nonetheless* can make paragraphs more stilted.)
+- Even if you're having trouble hitting the right tone, make sure you're
+  communicating useful information in a clear and direct way; that's the most
+  important part.
+
+<a id="tone--politeness"></a>
+
+<a id="tone--politeness-and-use-of-please"></a>
+
+#### Politeness and use of *please*
+
+It's great to be polite, but using *please* in a set of instructions is
+overdoing the politeness.
+
+Recommended: To view the document, click
+**View**.
+
+Not recommended: To view the document,
+please click **View**.
+
+Recommended: For more information, see
+[link to other document].
+
+Not recommended: For more information,
+please see [link to other document].
+
+<a id="tone--examples"></a>
+
+#### Examples
+
+| Too informal | Just about right | Too formal |
+| --- | --- | --- |
+| Dude! This API is totally awesome! | This API lets you collect data about what your users like. | The API documented by this page may enable the acquisition of information pertaining to user preferences. |
+| Just like a certain pop star, this call gets your *telephone* number. The easy way to ask for someone's digits! | To get the user's phone number, call `user.phoneNumber.get`. | The telephone number can be retrieved by the developer via the simple expedient of using the `get` method on the `user` object's `phoneNumber` property. |
+| Then—BOOM—just garbage-collect, and you're golden. | To clean up, call the `collectGarbage` method. | Please note that completion of the task requires the following prerequisite: executing an automated memory management function. |
+
+---
+
+## Language and grammar
+
+<a id="abbreviations"></a>
+
+### Abbreviations
+
+*Source: <https://developers.google.com/style/abbreviations>*
+
+Abbreviations include acronyms, initialisms, shortened words, and
+contractions.
+
+In most contexts, the technical distinction between acronyms and initialisms
+isn't relevant; it's fine to use the word *acronym* to refer to both.
+
+- An acronym is formed from the first letters of words in a phrase, but is
+  pronounced as if it were a word itself:
+  - *NATO* for *North Atlantic Treaty Organization*
+  - *scuba* for *self-contained underwater breathing
+    apparatus*
+- An initialism is also formed from the first letters of words in a phrase,
+  but each letter is pronounced separately:
+  - *CIA* for *Central Intelligence Agency*
+  - *FYI* for *For Your Information*
+  - *PR* for *Public Relations*
+- A shortened word is just part of a word or phrase, sometimes with a
+  period at the end:
+  - *Dr.* for *doctor*
+  - *etc.* for *et cetera*
+  - *min* for *minutes*
+  - *CA* for *California*
+- [Contractions](#contractions) are discussed in a
+  separate page of this style guide.
+
+<a id="abbreviations--long-and-short-versions"></a>
+
+#### Long and short versions of a word
+
+Some words have a long version and a short version—for example:
+
+- *application* and *app*
+- *demonstration* and *demo*
+- *synchronize* and *sync*
+
+The short versions of the words are not abbreviations, and
+if you use them, you don't need to put a period after them.
+
+If you're not sure whether a word is an abbreviation or just a short version
+of a longer word, look in our list of [resources](#style--editorial-resources).
+If that doesn't settle the issue, use the speaking test: if you speak the short
+version as a word (*This is a demo version of the product*), you can usually
+treat it as a word and not an abbreviation.
+
+<a id="abbreviations--creating-abbreviations"></a>
+
+#### When to use abbreviations
+
+Abbreviations are intended to save the writer and the reader time. If the reader has to
+think about an abbreviation, it can slow down their reading comprehension.
+
+<a id="abbreviations--general-dos-and-donts"></a>
+
+##### General dos and don'ts
+
+- Use standard acronyms and initialisms that will save the reader time.
+- Spell out abbreviations on first reference. For more information,
+  see [When to spell out a term](#abbreviations--spelling-out).
+- Avoid using abbreviations for terms that aren't related to the main topic of the document.
+
+In the following examples, the main topic of the document is the internet of things, so *low Earth orbit* should not be abbreviated.
+
+Recommended: The internet of things (IoT)
+service can even be used for connecting to sensors in low Earth orbit.
+
+Not recommended: The IoT (internet of things)
+service can even be used for connecting to sensors in LEO (low Earth orbit).
+
+- Be wary of using specialized abbreviations that your readers might not understand. For more information about when to use such language, see [Jargon](#jargon).
+
+<a id="abbreviations--spelling-out"></a>
+
+#### When to spell out a term
+
+In general, when an abbreviation is likely to be unfamiliar to the audience, spell it out upon
+first mention and include the abbreviation in parentheses immediately following.
+
+Recommended:  *Border Gateway Protocol* (*BGP*)
+
+For all
+subsequent mentions of the abbreviation, use the abbreviation by itself. If you use an abbreviation
+only once, include it only if you think the abbreviation is as commonly used as the spelled-out
+term. Otherwise, don't include the abbreviation.
+
+If the first mention of
+a term occurs in a [heading or title](#headings), you can use the abbreviation and
+then spell out the abbreviation in the first paragraph that follows the heading or title.
+
+When deciding to spell out a term, consider your audience. If your document is going to be
+translated, spelling out a term can provide important context for both human and machine translation.
+It can also be helpful for readers who aren't as familiar with English. If the
+majority of your audience is likely to recognize and understand the term, then you don't need to
+spell it out. For example, if you're writing documentation for developers that references an API,
+you don't need to spell out *application programming interface*. However, if you're explaining
+the general concept of an API to someone with no programming experience, spelling out the
+abbreviation can be helpful.
+
+In some cases, spelling out a term doesn't help the reader
+understand the term. For example, writing out *portable document format* doesn't help the
+reader understand what a *PDF* document is. In those cases, don't spell out the term.
+
+The following abbreviations rarely need to be spelled out:
+
+- AI
+- API
+- DVD
+- File formats such as PDF or XML
+- HTML
+- PC
+- RAM
+- REST
+- [Units of measurement](#units-of-measure--byte-units) such as
+  MB, MiB, GB, or GiB
+- URL
+- USB
+
+<a id="abbreviations--format-abbreviations"></a>
+
+##### Format abbreviation introductions
+
+When you spell out a term and include its abbreviation in parentheses, do the following:
+
+- Italicize both the spelled-out term and its abbreviation.
+
+In the following examples, the only difference is the italicization of *BGP*:
+
+Recommended: Establish *Border Gateway Protocol* (*BGP*) sessions using a router on the peer network.
+
+Not recommended: Establish *Border Gateway Protocol* (BGP) sessions using a router on the peer network.
+
+- Capitalize the spelled-out version of the abbreviation only if the long form is a proper noun or is conventionally capitalized.
+
+That is, don't capitalize the term only because the abbreviation includes capital letters.
+
+Recommended: data manipulation language (DML)
+
+Not recommended: Data Manipulation Language (DML)
+
+- Include the abbreviation in [link text](#cross-references--abbreviations).
+
+<a id="abbreviations--dont-use"></a>
+
+#### Abbreviations not to use
+
+Don't use *i.e.* or *e.g.*; instead, use *that is* or
+*for example*, respectively. For more information, see
+[e.g.](#word-list--eg) and
+[i.e.](#word-list--ie)
+
+It's [okay to use *etc.* in
+some circumstances](#word-list--etc), but it's best to use different phrasing in most lists.
+For more information, see
+[Comma-separated lists](#lists--comma-separated-lists).
+
+Don't use internet slang abbreviations such as
+[*tl;dr*](#word-list--tldr),
+[*ymmv*](#word-list--ymmv),
+[*RTFM*](#word-list--rtfm), or others. Write out what you
+mean in a non-figurative way.
+
+Use the most common form of a word. If the spelled-out word is common
+and easily understandable, use that rather than abbreviating. For example, write
+*approximately* instead of *approx.*
+
+Spell out shortened words or common symbols that are substitutions for words.
+
+Recommended: Updating the software made
+throughput 10 times faster.
+
+Not recommended: Updating the software made
+throughput 10x faster.
+
+<a id="abbreviations--periods"></a>
+
+#### Periods with abbreviations
+
+Follow these guidelines:
+
+- Don't use periods with acronyms or initialisms.
+- Put a period at the end of a shortened word, except for
+  [date and time](https://developers.google.com/style/dates) abbreviations.
+- If you write or say an abbreviation as a word (for example, *app* or
+  *sync*), don't put a period after it.
+- Don't use a period with an abbreviation for the name of a country, US
+  state, or the District of Columbia (DC).
+
+<a id="abbreviations--making-abbreviations-plural"></a>
+
+#### Plural abbreviations
+
+For guidance about how to make abbreviations plural, see [Pluralization](#pluralization--making-abbreviations-plural).
+
+<a id="abbreviations--verbs"></a>
+
+#### Abbreviations as verbs
+
+Don't use acronyms, initialisms, or shortened words as verbs.
+
+Recommended: Use SSH to
+log in to your remote shell.
+
+Not recommended: Then ssh
+into your remote shell.
+
+<a id="abbreviations--articles"></a>
+
+#### Indefinite articles before abbreviations
+
+Whether to use *a* or *an* before a term depends on the pronunciation of the term:
+use *a* before any consonant sound and *an* before any vowel sound. Pronunciation of
+abbreviations can vary, so in general, base your decision on the pronunciation that's most common
+for your audience.
+
+In particular, our word list includes preferences for
+"[a SQL](#word-list--sql)", "[a FHIR](#word-list--fhir)",
+and "[an SAP](#word-list--sap)".
+
+For more information about articles, see [Articles](#articles).
+
+---
+
+<a id="voice"></a>
+
+### Active voice
+
+*Source: <https://developers.google.com/style/voice>*
+
+In general, use active voice (in which the grammatical subject of the
+sentence is the person or thing performing the action) instead of passive voice
+(in which the grammatical subject of the sentence is the person or thing being
+acted upon), although there are exceptions. Make clear who's performing the
+action.
+
+In passive voice, it's easy to neglect to indicate who
+or what is performing a particular action. In this kind of construction, it's
+often hard for readers to figure out who's supposed to do something (such as the
+reader, the computer, the server, an end user, or a visitor to a web page).
+
+Recommended: Send a query to the service.
+The server sends an acknowledgment.
+
+Not recommended: The service is queried,
+and an acknowledgment is sent.
+
+It's possible to indicate who's performing the action in passive voice (using
+*by*), but the resulting prose is generally not as good as if you were to recast
+the sentence as active voice. So whenever possible, make the doer the subject of
+the sentence.
+
+Recommended: Send a query to the service.
+The server sends an acknowledgment.
+
+Not recommended: The service is queried by
+you, and an acknowledgment is sent by the server.
+
+For more information, see
+[Active voice vs. passive voice](https://developers.google.com/tech-writing/one/active-voice)
+in Google's Technical Writing One guide.
+
+<a id="voice--exceptions"></a>
+
+#### Exceptions
+
+In certain cases, it's okay to use passive voice. For example, passive can be
+okay in the following instances:
+
+- To emphasize an object over an action.
+
+Recommended: The file is saved.
+
+- To de-emphasize a subject or actor.
+
+Recommended: Over 50 conflicts were
+found in the file.
+
+Not recommended: You created over 50
+conflicts in the file.
+
+- If your readers don't need to know who's responsible for the action.
+
+Recommended: The database was purged
+in January.
+
+---
+
+<a id="anthropomorphism"></a>
+
+### Anthropomorphism
+
+*Source: <https://developers.google.com/style/anthropomorphism>*
+
+Don't attribute human qualities to software or hardware.
+
+Anthropomorphism is a category of figurative language, which is less precise and is often harder
+to understand and translate than direct language. For more information, see
+[Write for a global audience](#translation).
+
+Recommended: A Delimiter object specifies
+where to split a string.
+
+Not recommended: A Delimiter object tells
+the splitter where a string should be broken.
+
+Recommended: The PC detects a new
+device.
+
+Not recommended: The PC sees a new
+device.
+
+---
+
+<a id="articles"></a>
+
+### Articles (a, an, the)
+
+*Source: <https://developers.google.com/style/articles>*
+
+For ease of comprehension and translation, include definite and indefinite
+articles (*a*, *an*, and *the*) in your writing. Don't skip
+articles for brevity, including in headings and titles.
+
+Recommended: Create a VM instance
+
+Not recommended: Create VM instance
+
+For more information about using standard English word order and about writing
+for a global audience in general, see
+[Write for a global audience](#translation).
+
+For more information about writing clear headings and titles, see
+[Headings and titles](#headings).
+
+For information about using articles before product names, see
+[Articles before product names](#product-names--the-with-names).
+
+For information about using *a* or *an* before an abbreviation when the pronunciation
+of the abbreviation can vary, see
+[Indefinite articles before abbreviations](#abbreviations--articles).
+
+---
+
+<a id="capitalization"></a>
+
+### Capitalization
+
+*Source: <https://developers.google.com/style/capitalization>*
+
+Follow the standard [capitalization rules](https://owl.purdue.edu/owl/general_writing/mechanics/help_with_capitals.html) for American English. Additionally,
+do the following:
+
+- Don't use unnecessary capitalization; before you capitalize a word, think
+  about why (and whether) it should be capitalized.
+- Don't rely on a difference in capitalization to convey meaning. For example,
+  although people who are familiar with Kubernetes probably understand that a
+  capitalized *Pod* is a Kubernetes unit, and a lowercase *pod* is
+  any other kind of pod, that distinction is likely lost on many casual
+  readers or those who are new to the domain.
+- Don't use all-uppercase, except in the following contexts: in official
+  names, in [abbreviations](#abbreviations) that are always
+  written in all-caps, or when referring to code that uses all-caps.
+- Don't use
+  [camel
+  case](https://en.wikipedia.org/wiki/Camel_case), except in official names or when referring to code that uses camel
+  case.
+
+For information about how to capitalize specific words, see the
+[word list](#word-list).
+
+<a id="capitalization--capitalize-product-names"></a>
+
+#### Capitalize product names
+
+For information about how to capitalize product names, see
+[Product names](#product-names).
+
+<a id="capitalization--capitalization-in-titles-and-headings"></a>
+
+#### Capitalization in titles and headings
+
+In [document titles and headings](#headings), use sentence case. That is,
+capitalize only the first word in the title, the first word in a subheading after a colon, and any
+proper nouns or other terms that are always capitalized a certain way.
+
+Even though you're using sentence case, don't put a period at the end of a title or
+heading.
+
+<a id="capitalization--capitalization-in-references-to-titles-and-headings"></a>
+
+##### Capitalization in references to titles and headings
+
+In references to any title or heading from a document that follows this guide, use sentence case
+even if the title or heading itself uses title case. That way, when the title or heading is
+eventually updated to sentence case, the reference will match.
+
+When you reference the title of any work or source that doesn't follow this guide, retain the
+original capitalization.
+
+For more information about internal and external references, see
+[Cross-references and linking](#cross-references).
+
+For more information about formatting references to third-party sources,
+see [HTML and semantic tagging](#semantic-tagging).
+
+<a id="capitalization--capitalization-and-colons"></a>
+
+#### Capitalization and colons
+
+Use a lowercase letter to begin the first word of the text immediately
+following a colon, unless the text is one of the following:
+
+- A proper noun (*Open source software: Hadoop*)
+- A heading; see also [Capitalization in titles
+  and headings](#capitalization--capitalization-in-titles-and-headings)
+- A quotation (*Arthurian wit: "Bring me yon sworde"*)
+- Text that follows a label such as *Caution* or *Note*
+
+<a id="capitalization--capitalization-and-figures"></a>
+
+#### Capitalization and figures
+
+Use sentence case for captions. Use sentence case for labels, callouts, and
+other text in images and diagrams.
+
+<a id="capitalization--capitalization-in-glossaries-and-indexes"></a>
+
+#### Capitalization in glossaries and indexes
+
+Use lowercase for glossary and index terms unless the term is a proper noun
+or has another reason to require capitalization.
+
+Use sentence case for glossary definitions.
+
+<a id="capitalization--capitalization-and-hyphens"></a>
+
+<a id="capitalization--capitalization-and-hyphenated-words"></a>
+
+#### Capitalization and hyphenated words
+
+When a hyphenated word is the first word in a sentence or in a heading,
+capitalize only the first element in the word, unless a subsequent element is a
+proper noun or proper adjective.
+
+<a id="capitalization--capitalization-in-lists"></a>
+
+#### Capitalization in lists
+
+Use sentence case for items in all types of lists. For more information, see
+[Capitalization and end punctuation](#lists--capitalization-and-end-punctuation).
+
+<a id="capitalization--capitalization-for-tables-in-text"></a>
+
+#### Capitalization for tables in text
+
+Use sentence case for all the elements in a table: contents, headings,
+labels, and captions.
+
+<a id="capitalization--special-capitalization-style-names"></a>
+
+#### Special capitalization style names
+
+Don't use a casing style name, such as *camel case* or *snake case*, to describe a
+casing style. These names don't localize well and they aren't standardized. Instead, explain what
+the requirements are and provide an example.
+
+Recommended: Enter the value for the
+`attribute` field in the format where there are no spaces between words and the
+first letter of each word is capitalized—for example, `AssertionAccount`.
+
+---
+
+<a id="contractions"></a>
+
+### Contractions
+
+*Source: <https://developers.google.com/style/contractions>*
+
+In general, we write our documentation in an [informal tone](#tone), so we
+recommend using common two-word contractions such as *you're*, *don't*, and
+*there's*.
+
+<a id="contractions--negation-contractions"></a>
+
+#### Negation contractions
+
+In particular, we recommend using negation contractions such as *isn't*, *don't*, and
+*can't*. It's easy for a reader to miss the word *not* when they're scanning, whereas
+it's harder to misread *don't* as *do*.
+
+If you need to emphasize the negative, you can use text formatting such as `is
+<em>not</em>`, which renders as "is *not*." But in most cases, you don't
+need emphasis to make your point clear.
+
+<a id="contractions--contractions-to-avoid"></a>
+
+#### Contractions to avoid
+
+Don't make up nonstandard contractions such as *guides're* or *browser's* (where
+*'s* means *is*).
+
+Don't use three-word contractions such as *mightn't've*.
+
+---
+
+<a id="pluralization"></a>
+
+### Pluralization
+
+*Source: <https://developers.google.com/style/pluralization>*
+
+In general, follow the standard rules for pluralization in US English and use the regular
+plural form of a word in most cases. Avoid using *'s* to form a plural to avoid confusing
+a plural with a possessive or contraction.
+
+For more information, see [Contractions](#contractions) and [Possessives](#possessives).
+
+<a id="pluralization--singular-and-plural"></a>
+
+#### Singular and plural
+
+For sentences with long or complex subjects, make sure to use either a plural or singular
+appropriately.
+
+Recommended: Confirm that the number of
+entries listed in the directory is accurate.
+
+Recommended: The workloads with the
+`app: backend` label represent the traffic source.
+
+Not recommended: The efficiency of
+algorithms that process data sets depend on memory allocation.
+
+For sentences with more than one subject being connected by *and* or *or*, make sure
+to use either a plural and singular appropriately.
+
+Recommended: The request payload and
+header information are logged for debugging.
+
+Recommended: Either the API keys or
+service account wasn't authenticated.
+
+Not recommended: User authentication and
+authorization is processed and handled by the security module.
+
+For consistent style, use a plural after *one or more*, not a singular.
+
+Because *one or more* can express the possibility of one or more outcomes, it's sometimes
+helpful to reword the sentence for clarity.
+
+Recommended: If one or more tests fail, a
+system warning is triggered.
+
+Recommended: If any one test fails, a
+system warning is triggered.
+
+See also [Plurals in parentheses](#pluralization--plurals-in-parentheses).
+
+Use a singular after *more than one*, not a plural.
+
+Recommended: You can create more than one
+instance at a time.
+
+<a id="pluralization--making-abbreviations-plural"></a>
+
+#### Plural abbreviations
+
+In general, treat acronyms, initialisms, and other abbreviations as regular words when making
+them plural. Avoid using *'s* to form the plural to help distinguish the plural form from the
+possessive. For more information, see [Possessives](#possessives).
+
+Recommended: APIs, SKEs, and IDEs
+
+Not recommended: API's, SKE's, and
+IDE's
+
+If the acronym, initialism, or abbreviation ends in *s*, *sh*, *ch*, or
+*x*, then add *es*—for example, *OSes*, *DISHes*, *DCCHes*, and
+*BMXes*.
+
+> [!NOTE]
+> **Note:** Forming a plural using *'s* can be ambiguous or confusing
+> to readers and might cause issues in translation. For more information, see [Write for a global audience](#translation).
+
+When spelling out a term, make sure that both the spelled-out term and abbreviation match,
+with both either being a plural or singular.
+
+Recommended: virtual machines (VMs)
+
+Not recommended: virtual machines (VM)
+
+When using numbers with units of measure, use the singular when spelling out the unit if the
+number is one. Otherwise, use the plural form for all other numbers, including zero,
+decimal numbers, and numbers greater than one.
+
+Recommended: 0 degrees
+
+Recommended: 0.5 degrees
+
+Recommended: 1 degree
+
+Recommended: 15 degrees
+
+Don't make an abbreviation plural when used as a unit with a number.
+
+Recommended: 64 GB
+
+Not recommended: 64 GBs
+
+> [!NOTE]
+> **Note:** Sometimes it can be helpful to spell out *-bit* or
+> *-byte* terms. However, in general, it's not necessary to spell out a unit when used in
+> combination with a specific number.
+
+Make sure to include a space, preferably a nonbreaking space, between the number and
+abbreviation. For more information, see [Spaces in units of
+measurement](#units-of-measure--spaces-in-units-of-measurement).
+
+For more information, see [Abbreviations](#abbreviations).
+
+<a id="pluralization--plural-product-and-feature-names"></a>
+
+#### Plural product and feature names
+
+In general, don't form a plural or possessive for the trademark of a product, feature, or
+company name. For more information, see [Use trademarks only as modifiers](#trademarks--use-trademarks-only-as-modifiers) and
+[Product, feature, and company
+names](#possessives--productu002c-featureu002c-and-company-names).
+
+In general, use singular class names. Don't manually make a singular class name plural. Doing
+so might cause issues in translation. Instead, add a plural noun after the class name.
+
+Recommended: `Intent` objects
+and `Activity` instances
+
+Not recommended: `Intent`s and
+`Activity`s
+
+Not recommended: `Intents` and
+`Activities`
+
+For more information, see [API reference code
+comments](#api-reference-comments).
+
+<a id="pluralization--plurals-in-parentheses"></a>
+
+#### Plurals in parentheses
+
+Don't put optional plurals in parentheses. Instead, use either a plural or singular
+construction and keep things consistent throughout your documentation. Choose what is most
+appropriate for your documentation and your audience. If it's important in a specific context to
+indicate both, use *one or more*.
+
+| Recommended | Not recommended |
+| --- | --- |
+| To find your API key, visit the **Credentials** page. | To find your API key(s), visit the **Credentials** page. |
+| The value of the parent depends on the values of its children. | The value of the parent depends on the value(s) of its child(ren). |
+| You can use a physical linecard, which can contain one or more ports. | You can use a physical linecard, which can contain port(s). |
+
+<a id="pluralization--plural-pronouns"></a>
+
+#### Plural pronouns
+
+For information about plural pronouns like *we*, *you*, and *they*, see [Pronouns](#pronouns) and [Second person and first
+person](#person).
+
+---
+
+<a id="possessives"></a>
+
+### Possessives
+
+*Source: <https://developers.google.com/style/possessives>*
+
+In general, to form a possessive, follow these guidelines.
+
+For singular nouns, including those that end in *s*, add *'s* to the end of the word.
+
+Recommended: Modify each vector's record.
+
+Recommended: Raise the storage class's quota.
+
+For plural nouns that end in *s*, add only an apostrophe to the end of the word.
+
+Recommended: Extend the models' capabilities.
+
+Not recommended: Extend the models's
+capabilities.
+
+For plural nouns that don't end in *s*, add *'s* to the end of the word.
+
+If a possessive seems awkward, rewrite the sentence to omit the possessive.
+
+Recommended: Analyze the business data.
+
+Not recommended: Analyze the businesses' data.
+
+Recommended: The rule that the Federal Trade
+Commission (FTC) issued.
+
+Not recommended: The Federal Trade
+Commission's (FTC's) rule.
+
+Avoid using *'s* to form a plural noun. For more information, see [Pluralization](#pluralization).
+
+<a id="possessives--productu002c-featureu002c-and-company-names"></a>
+
+#### Product, feature, and company names
+
+When describing function or performance, don't form a possessive from a
+feature name, product name, or trademark, regardless of who owns it. Instead,
+use the name as a modifier or rewrite to use a word like *of* to indicate
+the relationship.
+
+Recommended: You can use this template to
+monitor Google Search performance.
+
+Recommended: You can use this template to
+monitor the performance of Google Search.
+
+Not recommended: You can use this template to
+monitor Google Search's performance.
+
+To form the possessive of a company name, add *'s* to the end of the name. Don't form the
+possessive of a company name when using it as a trademark.
+
+Recommended: Google's new office is
+nearby.
+
+Not recommended: The capabilities of
+Google's Search are vast.
+
+For information about using trademarks as adjectives, not nouns, see
+[Trademarks](https://developers.google.com/style/trademarks#use-trademarks-as-adjectives).
+<a id="possessives--code-items"></a>
+
+#### Code items
+
+Don't form the possessive of a code item. Instead, form the possessive from the noun that
+follows the code item or rewrite to avoid the possessive form.
+
+Recommended: Compare the number to the
+`wordCount` method's return value.
+
+Recommended: Compare the number to the
+value returned by the `wordCount` method.
+
+Not recommended: Compare the number to
+`wordCount`'s return value.
+
+For more information, see [Grammatical treatment of code
+elements](#code-in-text--grammatical-treatment-of-code-elements).
+
+---
+
+<a id="prepositions"></a>
+
+### Prepositions
+
+*Source: <https://developers.google.com/style/prepositions>*
+
+There's no rule against placing a preposition at the end of a sentence.
+Place the preposition where it makes the most sense and makes the sentence easiest
+to read. Use prepositions as needed, even at the ends of sentences.
+
+Recommended: For details, see the client
+library documentation for the language you're interacting with.
+
+Not recommended: For details, see the
+client library documentation for the language with which you're interacting.
+
+Include prepositions that increase clarity, omit unnecessary prepositions,
+and don't clutter the sentence with too many prepositions.
+
+Recommended: The icon for the connector
+manager turns green within a few minutes, and the connector instance is
+displayed shortly after.
+
+For information about which preposition to use when referring to UI elements, see
+[UI elements and interaction](#ui-elements--prepositions).
+
+---
+
+<a id="tense"></a>
+
+### Present tense
+
+*Source: <https://developers.google.com/style/tense>*
+
+Use present tense for statements that describe general behavior that's not associated with
+a particular time.
+
+Recommended: Send a query to the service.
+The server sends an acknowledgment.
+
+Not recommended: Send a query to the
+service. The server will send an acknowledgment.
+
+However, it's fine to use future tense (*will*) to distinguish an action that will occur in
+the future.
+
+Recommended: Add the filename to the
+backup list. The file will be archived the next time the backup process runs.
+
+In the following example, future tense is appropriate because Pub/Sub sends
+messages asynchronously; messages are not received immediately by subscribers.
+
+Recommended: A message is sent that
+will notify any Pub/Sub subscribers.
+
+Not recommended: A message is sent
+that notifies any Pub/Sub subscribers.
+
+Don't use future tense to describe how a product or feature will work after the next release
+or update. For more information, see [Document future features](#future).
+
+Also avoid the hypothetical future *would*—for example:
+
+Recommended: If you send an unsubscribe
+message, the server removes you from the mailing list.
+
+Not recommended: You can send an
+unsubscribe message. The server would then remove you from the mailing list.
+
+---
+
+<a id="pronouns"></a>
+
+### Pronouns
+
+*Source: <https://developers.google.com/style/pronouns>*
+
+Ensure that a pronoun clearly refers
+to its antecedent (the noun that it's replacing).
+
+<a id="pronouns--ambiguous-pronoun-references"></a>
+
+#### Ambiguous pronoun references
+
+Avoid vague and confusing references between a pronoun and its antecedent.
+
+Recommended: If you type text in the
+field, the text doesn't change.
+
+Not recommended: If you type text in the
+field, it doesn't change.
+
+Recommended: The name of the function to
+execute in the given script. The name does not include parentheses or
+parameters.
+
+Not recommended: The name of the function
+to execute in the given script. It does not include parentheses or
+parameters.
+
+In many cases, it's best to follow a demonstrative pronoun (like *this* and *these*)
+with a noun.
+
+Recommended: Set this value to true.
+
+Not recommended: Set this to true.
+
+Recommended: These approaches are your
+best options.
+
+Not recommended: These are your best options.
+
+<a id="pronouns--gender-neutral-pronouns"></a>
+
+#### Gender-neutral pronouns
+
+Don't use gender-specific pronouns unless the person you're referring to is
+actually that gender.
+
+In particular, don't use *he*, *him*, *his*, *she*, or *her* as
+gender-neutral pronouns, and don't use *he/she* or *(s)he* or other such
+punctuational approaches. Instead, use the singular *they*.
+
+Singular *they* has been in use for a long time; for example, [Jane Austen used it](http://www.pemberley.com/janeinfo/austheir.html),
+and in 2015 the Washington Post [adopted
+it as part of their official style](https://www.washingtonpost.com/opinions/the-post-drops-the-mike--and-the-hyphen-in-e-mail/2015/12/04/ccd6e33a-98fa-11e5-8917-653b65c809eb_story.html).
+
+For more suggestions, see
+The Chicago Manual of Style, 16th edition, section 5.225,
+"Nine techniques for achieving gender neutrality."
+
+<a id="pronouns--optional-pronouns"></a>
+
+#### Optional pronouns
+
+To avoid ambiguity and clarify meaning in sentences, use optional pronouns such as
+*that* and *which*.
+
+| Recommended | Not recommended |
+| --- | --- |
+| Right-click the link that you want to open. | Right-click the link you want to open. |
+| You can use other option parameters, which are described in the following section. | You can use other option parameters, described in the following section. |
+
+For more information, see
+[Relative pronouns](#pronouns--relative-pronouns).
+
+<a id="pronouns--personal-pronouns"></a>
+
+#### Personal pronouns
+
+Avoid first-person pronouns (*I*, *we*, *us*, *our*, and *ours*) except
+in the following contexts:
+
+- The questions in FAQs.
+- A document whose author makes comments in the first person.
+- Using *we* to refer to your organization, after using your organization's
+  name. For example, "Example Pet Store recommends that you feed your aardvark
+  Standardized Aardvark Treats. We cannot guarantee the happiness of your aardvark
+  otherwise."
+
+Use the second-person pronoun (*you*) whenever possible. For more information about
+second person, see [Second person and first person](#person).
+
+<a id="pronouns--relative-pronouns"></a>
+
+#### Relative pronouns
+
+There are several relative pronouns. This section concerns only three of
+them: *that*, *which*, and *who*.
+
+*That* and *which* don't mean exactly the same thing, so don't substitute one
+for the other:
+
+- *That* introduces a restrictive clause. It isn't preceded by a comma.
+
+  Recommended: The echidna that has a
+  long snout is furry.
+
+  This sentence describes a particular echidna, the one that has a long
+  snout.
+- *Which* introduces a nonrestrictive clause and is preceded by a comma.
+
+  Recommended: The echidna, which has a
+  long snout, is furry.
+
+  This sentence describes all echidnas, and mentions in passing that they
+  all have long snouts.
+
+For more information about restrictive and nonrestrictive clauses and whether
+to use *that* or *which*, read
+[what
+Grammar Girl has to say on the subject](https://www.quickanddirtytips.com/articles/which-versus-that/).
+
+When you're referring to a person, you can use *who* instead of *that*. If you're not
+sure which pronoun is appropriate in your context, then it's generally OK to use *that*.
+
+You can use *whose* to refer to people, animals, and things. *Whose* is the possessive
+form of both *who* and *which*.
+
+Recommended: Examine the variables whose
+values are set at compile time.
+
+---
+
+<a id="person"></a>
+
+### Second person
+
+*Source: <https://developers.google.com/style/person>*
+
+<a id="person--address-the-reader-as-you"></a>
+
+#### Address the reader as *you*
+
+In general, address the reader of your documents
+using the second person instead of the first person: use *you* or *your*
+instead of *we*, *our*, or *us*.
+Assume that the reader is the person who's doing the
+tasks or making the decisions. Use the word *user* only to refer to the user
+of the software that your reader is developing.
+
+| Recommended | Not recommended |
+| --- | --- |
+| The following sections describe how you can create a website. | The following sections describe how we can create a website. |
+| Consider adding a description to your table. | Let's add a description to our table. |
+| This document shows you how to develop an app for your organization. | This document shows the user how to develop an app for their organization. |
+
+If you're telling the reader to do something, then use the imperative (the *you* is
+implied). For example:
+
+Recommended: Click **Submit**.
+
+It's OK to use the imperative in running text after you establish who is being addressed.
+However, consider whether the imperative text needs to be formatted as as a procedure.
+
+Recommended: You can obtain the IP address
+for the appliance from your network administrator. Store the address in a variable for future
+use in the runbook.
+
+Not recommended: To hold the backup data,
+create a storage bucket. In the Google Cloud console, go to the **Buckets** page. Click
+**Create bucket**.
+
+There are some situations in which using *you* might not be accurate or
+appropriate. Use the second person to address what the reader does, but use the
+third person for what the software or an end user does. For example, in API
+documentation, you can use the third person when you state facts about programming
+elements, but address the reader as *you* when you tell them what to do with
+them.
+
+<a id="person--use-first-person-plural-pronouns-carefully"></a>
+
+#### Use first-person plural pronouns carefully
+
+It's OK to use first-person plural pronouns (such as *we*, *our*, or *us*)
+to refer to the organization that's represented as the author of the document. However, ensure
+that the antecedent for the pronoun is clear.
+
+Recommended: Example Organization provides
+A and B, but we don't provide C and D.
+
+Recommended: For more information, contact
+our sales organization.
+
+Recommended: The example.org support team
+regularly reviews tickets. Expect to hear from us in 2-3 business days.
+
+<a id="person--address-your-audience-consistently"></a>
+
+#### Address your audience consistently
+
+It's important to identify who the *you* is that you're addressing
+(a developer? a sysadmin? someone else?) and to be consistent
+about that. Make it clear to the reader who you expect them to be (sometimes
+with an explicit *audience* sentence near the beginning of the document).
+
+---
+
+<a id="sentence-structure"></a>
+
+### Sentence structure
+
+*Source: <https://developers.google.com/style/sentence-structure>*
+
+If you want to tell the reader to do something, try to mention the circumstance, conditions, or
+goal before you provide the instruction. Mentioning the circumstance first lets the reader skip
+the instruction if it doesn't apply. For information about how to apply this guideline to
+procedural instructions, see [Procedures](#procedures).
+
+| Recommended | Not recommended |
+| --- | --- |
+| For more information, see [link to other document]. | See [link to other document] for more information. |
+| To delete the entire document, click **Delete**. | Click **Delete** if you want to delete the entire document. |
+| If your app is located in one of the following regions, using custom domains might add noticeable latency to responses: | Using custom domains might add noticeable latency to responses if your app is located in one of the following regions: |
+
+---
+
+<a id="reference-verbs"></a>
+
+### Verbs in reference documents
+
+*Source: <https://developers.google.com/style/reference-verbs>*
+
+When you're writing reference documentation for a method, phrase the main
+method description in terms of what the method does (*gets*, *lists*, *creates*,
+*searches*), rather than what the developer would use it to do (*get*, *list*,
+*create*, *search*).
+
+It's a subtle distinction that manifests mostly in whether the initial verb
+in the description has an *-s* at the end or not.
+
+Recommended: tasks.insert: Creates a new
+task on the specified task list.
+
+Not recommended: tasks.insert: Create a
+new task on the specified task list.
+
+For more information and examples, see the
+[Google Cloud API design guide](https://cloud.google.com/apis/design/documentation#method_description).
+
+---
+
+## Punctuation
+
+<a id="colons"></a>
+
+### Colons
+
+*Source: <https://developers.google.com/style/colons>*
+
+A colon indicates that closely-related information follows.
+
+For information about using colons with run-in headings, see
+[Description lists that use
+run-in headings](#lists--description-lists-that-use-run-in-headings).
+
+<a id="colons--phrase-preceding-colon"></a>
+
+<a id="colons--introductory-phrase-preceding-colon"></a>
+
+#### Introductory phrase preceding colon
+
+When a colon introduces a list, the text that precedes the colon must be able
+to stand alone as a complete sentence.
+
+Recommended: The fields are defined as
+follows:
+
+Not recommended: The fields are:
+
+<a id="colons--colon-in-sentence"></a>
+
+<a id="colons--colons-within-sentences"></a>
+
+#### Colons within sentences
+
+In general, the first word in the text that follows a colon should be in
+lowercase. For exceptions, see
+[Capitalization and colons](#capitalization--capitalization-and-colons).
+
+Recommended: Tone: concise,
+conversational, friendly, respectful
+
+Recommended: When you add or update
+content to an existing project, remember to take these steps: review the style
+guide, use checklists, enlist a fellow writer or an editor to copyedit your
+work, and request a developmental edit if you feel that it's warranted.
+
+<a id="colons--see-also"></a>
+
+#### See also
+
+For more information about how to punctuate introductory material, see the
+sections on [list introductions](#lists--intros) and [code-sample introductions](#code-samples--intros).
+
+For information about when it's better to use colons than dashes, see [Dashes](#dashes--colons).
+
+---
+
+<a id="commas"></a>
+
+### Commas
+
+*Source: <https://developers.google.com/style/commas>*
+
+Use commas to separate items in a series, and use commas to separate certain kinds of
+clauses.
+<a id="commas--serial-commas"></a>
+
+#### Serial commas
+
+In a series of three or more items, use a comma before the final *and* or
+*or* to avoid potentially changing the meaning of the sentence. This comma is called a serial
+comma or an Oxford comma.
+
+Recommended: Locations are divided into
+zones, regions, and multi-regions.
+
+Not recommended: Locations are divided into
+zones, regions and multi-regions.
+
+<a id="commas--introductory-phrases"></a>
+
+<a id="commas--commas-after-introductory-words-and-phrases"></a>
+
+#### Commas after introductory words and phrases
+
+In general, place a comma after an introductory word or phrase.
+
+Recommended: Finally, only groups that
+contain parameters appear in this list.
+
+Recommended: Based on the requirements of
+your game, you can implement this method to update game information.
+
+<a id="commas--independent-clauses"></a>
+
+<a id="commas--commas-separating-two-independent-clauses"></a>
+
+#### Commas separating two independent clauses
+
+When a coordinating conjunction (*and*, *but*, *or*,
+*nor*, *for*, *so*, or *yet*) separates two independent
+clauses, insert a comma after the first clause (before the conjunction) unless
+both clauses are very short.
+
+Recommended: The libraries make
+feed creation easier, and they ensure that only valid feeds are produced.
+
+Not recommended: The libraries make
+feed creation easier and they ensure that only valid feeds are produced.
+
+Recommended: Type your ID and click **OK**.
+
+Not recommended: Type your ID, and click
+**OK**.
+
+<a id="commas--independent-dependent-clauses"></a>
+
+<a id="commas--commas-separating-independent-from-dependent-clauses"></a>
+
+#### Commas separating independent from dependent clauses
+
+When an independent clause and a dependent clause are separated by a
+coordinating conjunction, insert a comma *only if* the sentence could
+be misunderstood without one.
+
+Recommended: Direct-access flags are
+plain variables and can be read directly.
+
+Not recommended: Direct-access flags are
+plain variables, and can be read directly.
+
+Recommended: The manager acknowledged the
+last team member who entered the room, and started the meeting.
+
+Not recommended: The manager acknowledged
+the last team member who entered the room and started the meeting.
+
+<a id="commas--other-clauses"></a>
+
+<a id="commas--set-off-other-kinds-of-clauses"></a>
+
+#### Set off other kinds of clauses
+
+It's often a good idea to set off certain kinds of clauses with a comma or
+other punctuation for clarity.
+
+A couple of specific places where commas are a good idea:
+
+- In general, put a comma before the word *which* at the start of a
+  nonrestrictive clause. For more information about this topic, see this guide's section on [relative pronouns](#pronouns--relative-pronouns) and Grammar
+  Girl's page on
+  [*which* versus *that*](https://www.quickanddirtytips.com/articles/which-versus-that/).
+- In general, put a semicolon or a period or a dash before a conjunctive
+  adverb, such as *otherwise*, *however*, or *therefore*, and put a comma after
+  the conjunctive adverb.
+
+In general, don't use a comma before the causal conjunction *because* unless it starts a nonrestrictive clause. For more information,
+see the *Chicago Manual of Style* Q&A entry on
+[using
+commas with *because*](https://www.chicagomanualofstyle.org/qanda/data/faq/topics/Commas/faq0018.html).
+
+| Recommended | Not recommended |
+| --- | --- |
+| Name of the group, which has a maximum length of 200 characters. | Name of the group which has a maximum length of 200 characters. |
+| The variable must have a value; otherwise, the server returns an error. | The variable must have a value otherwise the server returns an error. |
+| You can use the same key name in multiple backend services and backend buckets, because each set of keys is independent of the others. | You can use the same key name in multiple backend services and backend buckets because each set of keys is independent of the others. |
+
+<a id="commas--numbers"></a>
+
+<a id="commas--punctuate-numbers"></a>
+
+#### Punctuate numbers
+
+For information about punctuating numbers, see [Commas and decimal
+points in numbers](#numbers--commas-and-decimal-points-in-numbers).
+
+<a id="commas--punctuate-examples"></a>
+
+#### Punctuate examples
+
+For information about punctuating examples, see [Format examples](#format-examples).
+
+---
+
+<a id="dashes"></a>
+
+### Dashes
+
+*Source: <https://developers.google.com/style/dashes>*
+
+This page explains when to use em dashes. For information about hyphens, see the following:
+
+- [Hyphens](#hyphens)
+- [Ranges of numbers](#numbers--ranges-of-numbers)
+- [Ranges of numbers with units](#units-of-measure--ranges)
+
+<a id="dashes--dashes"></a>
+
+<a id="dashes--em-dashes"></a>
+
+#### Em dashes
+
+To indicate a break in the flow of a sentence—or an interruption—use an em
+dash, also known as a long dash. Don't put a space before or after it.
+
+You can type the em dash character in various ways:
+
+**HTML**
+    &mdash;
+
+**macOS**
+    Press `Option+Shift+hyphen`.
+
+**Linux desktop environment**
+    Enable the Compose key (instructions for doing that vary depending on
+    your flavor of Linux—for examples, see [Linux Keyboard Shortcuts For
+    Text Symbols](http://fsymbols.com/keyboard/linux/compose/)). After the Compose key is enabled, you can create an em dash
+    by typing the Compose key followed by three hyphens.
+    Alternatively, press `Control+Shift+U`. Let go of those keys, and then type
+    `2014`. Then press `Return`.
+    > [!NOTE]
+    > **Note**: These Linux options don't work if
+    > you're signed in to the Linux command line from a remote system using `ssh` or the
+    > like; you have to be in a Linux desktop environment.
+
+**Windows**
+    Turn num lock on, and then hold down the left `Alt` key and type `0151`
+    on the numeric keypad.
+
+Don't use an en dash (the shorter dash) or a hyphen in place of an em dash.
+The use of an en dash with spaces around it in place of
+an em dash is gradually becoming more common, but it's still not very widespread
+in the US in professional publishing; so far (as of early 2016), it's mostly
+used in Canada and a few other places. For now, only use the em dash.
+
+<a id="dashes--ndashes"></a>
+
+<a id="dashes--en-dashes"></a>
+
+#### En dashes
+
+Don't use. Instead, use a hyphen or the word *to*. For more information, see
+the following:
+
+- [Ranges of numbers with units](#units-of-measure--ranges)
+- [Range of numbers](#hyphens--number-range)
+
+<a id="dashes--colons"></a>
+
+<a id="dashes--colons-instead-of-dashes-in-description-lists"></a>
+
+#### Colons instead of dashes in description lists
+
+Another common but nonstandard construction is to use an em dash, an en dash, or a hyphen
+surrounded by spaces to separate an item and its description. Instead, use
+[a colon or a period](#lists--description-lists-that-use-run-in-headings).
+For a series of items, use
+[an HTML description list](#lists--description-lists) (`<dl>`).
+
+Recommended: Example: This is an
+example.
+
+Not recommended: Example - This is
+an example.
+
+Recommended: Appendix A: My first
+appendix
+
+Not recommended: Appendix A—My first
+appendix
+
+Recommended:
+
+```
+    <dl>
+      <dt>Example</dt>
+      <dd>This is an example.</dd>
+      <dt>Another example</dt>
+      <dd>This is another example.</dd>
+    </dl>
+    
+```
+
+---
+
+<a id="ellipses"></a>
+
+### Ellipses
+
+*Source: <https://developers.google.com/style/ellipses>*
+
+In general, don't use ellipses. An ellipsis is made up of
+three contiguous periods. Ellipses indicate the omission of part of a sentence, paragraph, or larger
+block of text where the omission is not pertinent to the understanding of the subject at
+hand.
+
+<a id="ellipses--suspension-points"></a>
+
+<a id="ellipses--ellipses-as-suspension-points"></a>
+
+#### Ellipses as suspension points
+
+When ellipses are used to indicate hesitation, they are called *suspension
+points*. Don't use ellipses this way in our documentation.
+
+Not recommended: The answer is ... wait
+for it ... that you shouldn't do this.
+
+<a id="ellipses--how-to-use-ellipses"></a>
+
+<a id="ellipses--ellipses-in-a-user-interface"></a>
+
+#### Ellipses in a user interface
+
+When ellipses appear in a user interface, exclude them from the
+documentation describing the user interface unless their omission could cause
+confusion. For example, if the text on the button in the UI reads **Save ...**,
+document it as *click **Save***.
+
+<a id="ellipses--ellipses-in-text"></a>
+
+#### Ellipses in text
+
+Don't use ellipses in your written documentation; omit any unnecessary
+information and include all necessary information.
+
+However, it's acceptable to use ellipses in quoted text (to replace a
+portion of the quoted text) except when they appear at the beginning or end of
+the text.
+
+Not recommended: My high school English
+teacher made me learn that Shakespeare quote about all the world being a stage
+and " ... all the men and women merely players."
+
+Not recommended: My high school English
+teacher made me learn that Shakespeare quote: "All the world's a stage, And all
+the men and women merely players ...."
+
+The previous example ended with four ellipsis points. The final
+ellipsis point is, in fact, a period. So when the material that you're omitting
+contains one or more sentence boundaries, use four dots instead of three.
+
+Recommended: My high school English
+teacher made me learn that Shakespeare quote: "All the world's a stage, ....
+And one man in his time plays many parts."
+
+<a id="ellipses--punctuation-and-spacing-of-ellipses"></a>
+
+#### Punctuation and spacing of ellipses
+
+Keep all three ellipsis points together. When creating an ellipsis,
+instead of the ellipsis character, use three periods in a row. Insert one space
+before and after the ellipsis unless a punctuation mark immediately follows the
+ellipsis; in this case, don't insert a space after the ellipsis.
+
+Recommended: You don't need to
+understand all the other Python code in there ... we'll explain it all in class.
+
+Also recommended: You don't need to
+understand all the other Python code in there ...; we'll explain it all in class.
+
+Not recommended: You don't need to
+understand all the other Python code in there...we'll explain it all in class.
+
+---
+
+<a id="hyphens"></a>
+
+### Hyphens
+
+*Source: <https://developers.google.com/style/hyphens>*
+
+Use a hyphen (-) when needed for clarity. A hyphen can separate parts of words to avoid
+misreadings, and it can combine terms when they should be read as a unit.
+
+<a id="hyphens--general-guidelines"></a>
+
+#### General guidelines
+
+Guidance for hyphenation isn't always straightforward because it depends on
+the following circumstances:
+
+- **Location**. For example, does a term precede a noun, or does it follow a
+  verb?
+- **Interpretation and readability**. Is a sentence ambiguous or unclear if a term is
+  not hyphenated?
+- **Convention**. For some terms, our guidance tells us to always hyphenate or
+  never hyphenate, even if the convention seems to contradict other guidance.
+
+In addition, there are many exceptions to general hyphenation guidance. If you're not
+sure whether to hyphenate a term, in addition to reviewing the guidelines on this page, check the
+following sources (in this order):
+
+1. The documentation that you're working with. If there's an established
+   convention for hyphenating a term in a particular documentation set, follow that
+   convention.
+2. The [word list](#word-list) in this style guide.
+3. The [Merriam-Webster
+   dictionary](https://www.merriam-webster.com/).
+
+As always, deviate from our guidance when it serves your readers. For
+more information, see [Break the rules](#style--rules).
+
+> [!NOTE]
+> **Note**: Don't use a hyphen (-) or a double
+> hyphen (--) in place of a dash (—). The dash is a distinct punctuation mark
+> that has different uses. For more information, see
+> [Dashes](#dashes).
+
+<a id="hyphens--prefixes"></a>
+
+#### Prefixes
+
+In general, don't use a hyphen between a prefix and the main noun.
+
+Recommended: *infrastructure*,
+*megabyte*, *metadata*, *preprocessing*, *pseudocode*,
+*semiconductor*
+
+<a id="hyphens--exceptions"></a>
+
+##### Exceptions
+
+Add a hyphen after a prefix in the following circumstances:
+
+- If the prefix is *self* or *cross*: *self-managing*,
+  *cross-region*
+- If the noun is capitalized or is a number: *non-Google*,
+  *post-2000*
+- To avoid confusion or difficulty in reading: *de-energize*, *intra-index*,
+  *re-mark*, *re-sign*
+- If the prefix is for a term that already has hyphens or spaces:
+  *un-Google-like*, *non-twentieth-century*
+- To be consistent within a document: *pre-processing*,
+  *post-processing*
+
+<a id="hyphens--the-non-prefix"></a>
+
+##### The *non* prefix
+
+The *non* prefix follows the same guidelines, but because it
+can easily form words that are hard to parse, it's often hyphenated. Use your
+best judgment, taking into account consistency within your documentation. The following
+recommendations show contrasting usages that you can use as examples.
+
+Recommended: *noncurrent*,
+*nonempty*, *noninteractive*, *nonpublic*
+
+Recommended: *non-existence*,
+*non-integer*, *non-key*, *non-managed*, *non-negative*
+
+When using *non* as a prefix, add a hyphen before hyphenated compound words.
+
+Recommended: *non-KSA-based*,
+*non-self-sustaining*
+<a id="hyphens--compounds"></a>
+
+#### Compounds
+
+A *compound* is a term that combines more than one word. Compounds can
+be *closed* as one word with no spaces, *open* with spaces between
+words, or hyphenated.
+
+<a id="hyphens--compound-nouns"></a>
+
+##### Compound nouns
+
+In general, write compound nouns in their closed (one-word, unhyphenated)
+form. If you see that
+[Merriam-Webster.com](https://www.merriam-webster.com/) uses the
+two-word or hyphenated form, but you see that the closed form is the
+predominant convention in your context or trending in that direction (as
+compounds often do), then use the closed form.
+
+Recommended: webpage
+
+Recommended: hostname
+
+Recommended: tradeoff
+
+Recommended: workaround
+
+<a id="hyphens--exceptions_1"></a>
+
+###### Exceptions
+
+Our [word list](#word-list) includes exceptions for
+well-established terms that commonly use a hyphen or a space, such as
+*multi-region* and *style sheet*. In some cases, we note that noun,
+verb, and adjective versions of a word are treated differently.
+
+<a id="hyphens--multiplied-units"></a>
+
+When the components of a unit of measurement are multiplied by each other,
+hyphenate them.
+
+Recommended: 5 vCPU-hours
+
+Recommended: 40 person-hours
+
+<a id="hyphens--compound-modifiers-before-a-noun"></a>
+
+##### Compound modifiers before a noun
+
+If needed for clarity, hyphenate compound modifiers that come before a noun.
+This guideline can be subjective. However, except as noted in
+this section, it's almost never wrong to hyphenate a compound before a
+noun to ensure clarity.
+
+Recommended: A well-designed app
+
+Recommended: Android-specific
+techniques
+
+Use a hyphen after *more* or *most* if you need to clarify what
+those words modify.
+
+Recommended: The most common scenario
+
+Recommended: Edge locations with
+more-reliable internet links
+
+In general, avoid writing compound modifiers that have more than two words.
+Instead, move some words after the noun. If you must use this type of
+compound, then use a hyphen between each word as needed for clarity.
+
+Recommended: test cases
+that are specific to the 2023 edition
+
+Recommended:
+cross-data-center replication
+
+Not recommended:
+edition-2023-specific test cases
+
+<a id="hyphens--numbers-and-units-of-measurement"></a>
+
+###### Numbers and units of measurement
+
+Hyphenate a number and a spelled-out unit of measurement when they combine to modify a
+noun.
+
+Recommended: a 64-bit system
+
+Recommended: 100,000-byte files
+
+Recommended: a five-minute wait
+
+Don't hyphenate if the unit of measurement is abbreviated unless the hyphen is needed for
+clarity. Instead, use a nonbreaking space (`&nbsp;`) between the number and unit of
+measurement.
+
+Recommended: `200&nbsp;GB
+disk` (200 GB disk)
+
+Recommended:
+`50&nbsp;Mbps connection` (50 Mbps connection)
+
+For more information, see [Units of measurement](#units-of-measure).
+
+<a id="hyphens--exceptions_2"></a>
+
+###### Exceptions
+
+Don't hyphenate adverbs that end in *-ly* except when needed for clarity.
+
+Recommended: Publicly available
+implementations
+
+Not recommended: Publicly-available
+implementations
+
+Don't use hyphens in compounds that are conventionally not hyphenated. Follow the
+guidance in the [word list](#word-list)
+or check the convention in the documentation that you're working with.
+
+Recommended: A managed
+instance group (MIG)
+
+Recommended: A machine
+learning model
+
+<a id="hyphens--compound-terms-after-a-verb"></a>
+
+##### Compound terms after a verb
+
+In general, you don't need to add a hyphen to a compound that follows a verb.
+
+Recommended: The app is well
+designed.
+
+Recommended: The logs are written
+in real time.
+
+Recommended: The product supports
+high availability.
+
+Recommended: The app uses techniques
+that are Android specific.
+
+Recommended: Customers can use
+the utility as is.
+
+Recommended: Get profile information
+for the currently authorized user.
+
+<a id="hyphens--exceptions_3"></a>
+
+###### Exceptions
+
+Some compound terms are always hyphenated, even if they follow a verb. To
+check, look the term up in the [word list](#word-list). If it isn't in the
+list, check the [Merriam-Webster dictionary](https://www.merriam-webster.com/).
+As always, follow the convention in the documentation that you're working with.
+
+Recommended: You can deploy the app
+on-premises.
+
+Recommended: The docs describe how
+to create an add-on.
+
+Recommended: The utility works
+with apps that are cloud-based and cloud-adjacent.
+
+Recommended: This page is
+customer-facing.
+
+Recommended: The app is designed
+to be user-friendly.
+
+Recommended: The goal is to produce
+an experience that's game-like.
+
+<a id="hyphens--number-range"></a>
+
+<a id="hyphens--range-of-numbers"></a>
+
+#### Range of numbers
+
+Use a hyphen, not an en dash (`&ndash;`),
+to indicate a range of numbers. If a hyphen introduces ambiguity, use words such as
+*from*, *to*, and *through* for clarity. Don't mix hyphens with words.
+For information about how to represent a range of numbers that includes units, see
+[Ranges of numbers with units](#units-of-measure--ranges).
+
+Recommended: 8-20 files
+
+Recommended: 5-10 minutes
+
+Recommended: from 8 to 20 files
+
+Not recommended: from 8-20 files
+
+<a id="hyphens--spaces-around-hyphens"></a>
+
+#### Spaces around hyphens
+
+Never place a space on either side of a hyphen except when using a
+[suspended hyphen](#hyphens--suspended-hyphens), in which case you can leave a space after
+(but not before) the hyphen.
+
+<a id="hyphens--suspended-hyphens"></a>
+
+#### Suspended hyphens
+
+When two or more compound modifiers have a common base, you can keep the
+hyphens but leave out the base for all except the last modifier. In the
+following examples, the base is *hour*.
+
+Recommended: You can set up the system to
+scan for new files at one- or two-hour intervals.
+
+Recommended: You can set up the system to
+scan for new files at one-, two-, or three-hour intervals.
+
+---
+
+<a id="parentheses"></a>
+
+### Parentheses
+
+*Source: <https://developers.google.com/style/parentheses>*
+
+Some of us love to use parentheses. Unfortunately, some readers ignore
+anything that appears in parentheses, so don't put important information in
+parentheses if you can help it.
+
+Even for less important information, whenever you're inclined to use
+parentheses, consider whether they're necessary. Sometimes they are; however,
+the sentence or paragraph might work just as well if you remove the
+parentheses and set off the phrase or sentence by using commas, dashes, semicolons,
+or periods.
+
+If you need to include parentheses in the middle of a sentence, keep the parenthetical thought
+short. Otherwise, consider using two sentences.
+
+> [!NOTE]
+> **Note**: If a full standalone sentence appears inside
+> parentheses, the period also goes inside the parentheses, not outside.
+
+Recommended: Enter a name for the instance—for example, `my-instance-99`.
+
+Recommended: Enter a six-digit hex number (for example, `228B22`), and then click **OK**.
+
+Recommended: Enter a six-digit hex number, and then click **OK**. For example, if you want the color forest
+green, enter `228B22`.
+
+Not recommended: Enter a name for the instance (for example, `my-instance-99`).
+
+Not recommended: Enter a six-digit hex number (for example, if you want the color forest green, enter
+`228B22`), and then click **OK**.
+
+Don't use parentheses to indicate optional plurals. For more information, see [Plurals in parentheses](#pluralization--plurals-in-parentheses).
+
+---
+
+<a id="periods"></a>
+
+### Periods and end punctuation
+
+*Source: <https://developers.google.com/style/periods>*
+
+End a complete sentence with a period, unless it's a question. There are
+exceptions for working in lists.
+
+<a id="periods--periods-with-lists"></a>
+
+#### Periods with lists
+
+Whether to end a list item with a period depends on several factors, including
+the kind of list that the item appears in.
+
+For details about how to use periods in lists, see the
+[Capitalization and end punctuation](#lists--capitalization-and-end-punctuation)
+section of the "Lists" page.
+
+<a id="periods--periods-with-URLs"></a>
+
+<a id="periods--periods-with-urls"></a>
+
+#### Periods with URLs
+
+When a period immediately follows a URL or a file path, it can be hard to
+tell whether the period is part of the URL.
+
+To indicate that the punctuating period isn't part of the URL, try one of the
+following techniques:
+
+- Whenever possible, avoid putting [URLs in text](#cross-references--urls).
+- Rewrite the sentence so that the URL isn't at the end of the sentence.
+- Put the URL on a separate line from the text, omitting the final period.
+
+If the URL is a link, it generally looks different from the surrounding text. For
+example, in most browsers, link text is blue by default. This formatting helps
+distinguish the URL from the period.
+
+Recommended:
+
+We use your feedback to improve the Animals API, in accordance with Example
+Pet Store's Privacy Policy:
+
+http://www.examplepetstore.com/privacy/
+
+Not recommended:
+
+We use your feedback to improve the Animals API, in accordance with Example
+Pet Store's Privacy Policy at http://www.examplepetstore.com/privacy/.
+
+When you do put a period after a URL, don't leave any space between the last character of
+the URL and the period.
+
+<a id="periods--periods-with-quotation-marks"></a>
+
+#### Periods with quotation marks
+
+When a sentence ends with material inside quotation marks, place the period
+inside the quotation marks even if the period isn't part of the material inside
+the quotation marks. An exception to this guideline applies if you're using quotation marks around
+a keyword or other literal string. For more information, see
+[Commas and periods with quotation marks](#quotation-marks--commas-and-periods-with-quotation-marks).
+
+Recommended: ... you might say "Fixed typo."
+
+If the material inside the quotation marks ends with a question mark or an
+exclamation point, don't use a period.
+
+Recommended: Children always ask "Why?"
+
+<a id="periods--periods-with-parentheses"></a>
+
+#### Periods with parentheses
+
+If the last part of a sentence is contained inside parentheses, put the
+period after the closing parenthesis.
+
+If the parentheses contain a complete sentence, put the period inside
+the parentheses.
+
+Recommended: Your application could show
+a notification when a relevant file or folder has changed (even if that change
+occurs while your application isn't running).
+
+Recommended: App Engine applications are
+easy to create, easy to maintain, and easy to scale. (With App Engine, there are
+no servers for you to maintain.)
+
+For more information, see [Parentheses](#parentheses).
+
+<a id="periods--periods-with-headings"></a>
+
+<a id="periods--periods_with_headings"></a>
+
+#### Periods with headings
+
+Don't end headings with periods.
+
+For more information, see [Headings](#headings).
+
+<a id="periods--periods-with-numbers"></a>
+
+#### Periods with numbers
+
+Use a period to represent a decimal point. (Using a comma to separate the decimal part
+of a number is the editorial custom in some countries, but not in the US.)
+
+For more information, see [Numbers](#numbers).
+
+<a id="periods--periods-with-captions"></a>
+
+#### Periods with captions
+
+See [Figure captions](#images--figure-captions).
+
+<a id="periods--periods-with-alt-text"></a>
+
+#### Periods with alt text
+
+See [Alt text](#images--alt-text).
+
+<a id="periods--periods-with-abbreviations"></a>
+
+#### Periods with abbreviations
+
+Put a period after a shortened word.
+
+Don't put periods after the letters of an acronym or initialism.
+
+For more information, see [Abbreviations](#abbreviations).
+
+<a id="periods--sentence-spacing"></a>
+
+<a id="periods--spaces-between-sentences"></a>
+
+#### Spaces between sentences
+
+Leave only one space between sentences.
+
+<a id="periods--exclamation-points"></a>
+
+#### Exclamation points
+
+In general, avoid exclamation points. They can appear unprofessional, alarming, or translate poorly into other languages. If you need to use an exclamation point, see the following guidance by content type:
+
+- **Concept and reference docs.** Never use exclamation points. These sections should maintain a neutral, objective tone.
+- **Procedural topics.** Avoid exclamation points. Use periods for completion steps—for example, "The VM is created."
+- **Blog posts.** Exclamation points are acceptable to convey enthusiasm but shouldn't be used in every paragraph.
+
+When exclamation points are acceptable:
+
+- **Code examples.** Use when required by syntax—for example, the `!=` operator.
+- **System literals.** Use when an exclamation point is part of a specific error code or log message that must be matched exactly.
+- **Tutorials and learning modules.** Use sparingly to mark major milestones or achievements—for example, "Congratulations! You've completed the setup."
+
+> [!NOTE]
+> **Note:** For translation impact, be conservative with exclamation points in global documentation. In some languages, such as Japanese or Korean, exclamation marks can come across as overly emphatic or even shouting, which can alienate the reader.
+
+---
+
+<a id="quotation-marks"></a>
+
+### Quotation marks
+
+*Source: <https://developers.google.com/style/quotation-marks>*
+
+Use straight double quotation marks and apostrophes.
+
+<a id="quotation-marks--when-to-use-quotation-marks"></a>
+
+#### When to use quotation marks
+
+In technical writing, we don't use quotation marks much, aside from instances of code.
+
+Generally, you can use quotation marks for
+titles of shorter works such as articles or episodes in a web series, unless
+they're part of a link. For more information, see [cross-references](#cross-references).
+
+For most titles that are full-length works, we use italics.
+
+For examples of when to use quotation marks in regular text, see the following table:
+
+| Guidance | Example |
+| --- | --- |
+| Referring to a section of a larger document or piece, if you can't link to the section directly. | The technique is described in the section "Deploying containers" of the [Containers overview](https://www.youtube.com) video. |
+| Referring to the title of a parent document when you're already linking to a section | The [machine learning (ML) workflow section](https://cloud.google.com/vertex-ai/docs/start/introduction-unified-platform#ml-workflow) of "Introduction to Vertex AI" describes the machine learning workflow for Vertex AI. |
+| Directly citing a person or quoting a slogan or motto. | Martin Fowler has said, "We are still learning the techniques to write software effectively." |
+| Using a term metaphorically, but only if it's not an established usage in the domain. | This configuration forms an "island" within the network that is not connected to the external network. |
+
+For more information, see [Text-formatting summary](#text-formatting).
+<a id="quotation-marks--commaquote"></a>
+
+<a id="quotation-marks--commas-and-periods-with-quotation-marks"></a>
+
+#### Commas and periods with quotation marks
+
+Commas and periods go inside quotation marks.
+
+Recommended: See the section
+titled "Care and feeding of the emu."
+
+Not recommended: See the section titled "Care
+and feeding of the emu".
+
+**Exception**: When you put a keyword or other literal string in quotation
+marks, put any other punctuation outside the quotation marks. In those cases,
+the quotation marks indicate an exact literal string, so don't add anything
+extraneous inside the quotation marks. However, in general, don't put quotation marks
+around an item that's in code font, unless the quotation marks are part of the
+item.
+
+Recommended: If you enter `escape`,
+the program crashes.
+
+Acceptable: If you enter "escape", the program
+crashes.
+
+Not recommended: If you enter "escape," the
+program crashes.
+
+<a id="quotation-marks--curly-and-straight-quotation-marks"></a>
+
+<a id="quotation-marks--straight-and-curly-quotation-marks"></a>
+
+#### Straight and curly quotation marks
+
+Most typefaces support two forms of quotation marks and apostrophes:
+straight marks and curly, or typographic, marks. Some tools, like
+Google Docs, automatically convert straight quotation marks and
+apostrophes to the curly versions as you type. However, our guidance is
+to always use straight quotation marks and straight apostrophes in developer documentation, for
+the following reasons:
+
+- It makes writing documents easier.
+  - Code *requires* straight marks, so it's simpler to use straight marks everywhere
+    in developer documentation than to use them in code but not in text.
+  - Tools that automatically change straight marks to curly marks (such as word processors)
+    often make mistakes.
+  - Humans who manually type curly marks also often make mistakes.
+  - Manually typing curly marks can be difficult on some platforms.
+- It makes reviewing documents easier.
+  - When you're proofreading a document, it can be hard to see whether marks are straight or
+    curly, and which direction they point in.
+
+In the following examples, the first example uses straight quotation marks and the second example
+uses curly quotation marks:
+
+Recommended: The section's title is "Care
+and feeding of the emu."
+
+Not recommended: The section’s title
+is “Care and feeding of the emu.”
+
+<a id="quotation-marks--single-quotation-marks"></a>
+
+#### Single quotation marks
+
+The only times to use single quotation marks in our documentation are the following:
+
+- In code examples, in languages that use single quotation marks.
+- When nesting a quotation inside another quotation.
+
+In the latter case, put the primary speaker's quote in double quotation marks and the quote inside
+the primary speaker's quote in single quotation marks.
+
+Recommended: She said, "I heard
+him shout 'Help,' and saw him floundering in the water."
+
+Not recommended: She said, 'I heard him shout
+"Help", and saw him floundering in the water'.
+
+> [!NOTE]
+> **Note**: For information about how to use quotation marks with links, see
+> [Quotation marks and italics](https://developers.google.com/style/cross-references#quotation-marks-italics).
+
+---
+
+<a id="semicolons"></a>
+
+### Semicolons
+
+*Source: <https://developers.google.com/style/semicolons>*
+
+If possible, avoid using semicolons. In a few cases, a semicolon is preferred:
+
+- When joining two closely related independent clauses where a period or a comma is not as
+  effective.
+
+  Recommended: You can easily test
+  compatibility by computing the centroid; if it is on the opposite side of the
+  planet, reverse the order of your vertices.
+- When preceding a conjunctive adverb (like *therefore*) or a phrase
+  (like *that is*) that joins two independent clauses.
+
+  Recommended: This setup places the
+  head-tracked node below the Main Camera; therefore, only the stereo cameras are
+  affected by the user's head motion.
+
+  Recommended: The URL from which a video
+  ad loads; that is, the URL to use to fetch that video ad.
+- When separating a series of long or complex items that contain their own punctuation.
+
+  Recommended: If you don't have time,
+  then focus on the improvements that will have the greatest benefit: what matters most
+  to your users; what is most important to fix; and what is easy or feasible to
+  fix in the available time.
+
+  Recommended: Review your document one
+  more time, checking for the following: present tense and active voice; typos,
+  punctuation, and grammar; and whether you can shorten anything.
+
+  Notice that in the final example, the second item in the list is itself a list.
+
+---
+
+<a id="slashes"></a>
+
+### Slashes
+
+*Source: <https://developers.google.com/style/slashes>*
+
+Avoid using slashes, except in code.
+
+<a id="slashes--slashes-dates"></a>
+
+<a id="slashes--slashes-with-dates"></a>
+
+#### Slashes with dates
+
+Don't use date formats that rely on slashes.
+
+For information about how to write dates, see
+[Dates and times](#dates-times).
+
+<a id="slashes--slashes-alternatives"></a>
+
+<a id="slashes--slashes-with-alternatives"></a>
+
+#### Slashes with alternatives
+
+Don't use slashes to separate alternatives.
+
+Recommended: For example, a disaster
+relief map is not subject to the usage limits even if it has been developed and
+is hosted by a commercial entity.
+
+Recommended: For example, a disaster
+relief map is not subject to the usage limits even if it has been developed or
+is hosted by a commercial entity.
+
+Not recommended: For example, a disaster
+relief map is not subject to the usage limits even if it has been
+developed/hosted by a commercial entity.
+
+  
+
+Recommended: Call this method five or six
+times.
+
+Not recommended: Call this method 5/6
+times.
+
+<a id="slashes--and-or"></a>
+
+##### And/or
+
+Often, *and* implies *or*, so you don't need to write both words.
+If you need to specify both in your content, avoid writing *and/or* except
+when space is limited, such as in tables.
+
+Recommended: You can view
+and edit your own data.
+
+Not recommended: You can
+view and/or edit your own data.
+
+  
+
+Recommended: You can
+export raw events, processed events, or both.
+
+Not recommended: You can
+export raw and/or processed events.
+
+<a id="slashes--slashes-file-paths-urls"></a>
+
+<a id="slashes--slashes-with-file-paths-and-urls"></a>
+
+#### Slashes with file paths and URLs
+
+Use forward slashes, as appropriate, in computer file paths and URLs.
+
+> [!NOTE]
+> **Note**: If you're documenting a Windows path, use backslashes.
+
+Recommended:
+https://developers.google.com/cardboard/
+
+Where very long URLs extend beyond a line, add a line break immediately after
+a slash. Don't ever insert an extraneous hyphen into a URL to break it between two lines.
+
+Recommended:
+https://developers.google.com/
+  
+cardboard/
+
+<a id="slashes--slashes-fractions"></a>
+
+<a id="slashes--slashes-with-fractions"></a>
+
+#### Slashes with fractions
+
+Don't use slashes with fractions because they can be ambiguous.
+
+In the following example, 3/4 could be interpreted either as three-quarters
+or as stating that 4 is an alternative to 3.
+
+Recommended: ¾
+
+Recommended: 0.75
+
+Recommended: 75%
+
+Not recommended: 3/4
+
+<a id="slashes--slashes-abbreviations"></a>
+
+<a id="slashes--slashes-with-abbreviations"></a>
+
+#### Slashes with abbreviations
+
+Don't use abbreviations that rely on slashes. Instead, spell the words out.
+
+Recommended: care of, with
+
+Not recommended: c/o, w/
+
+---
+
+## Formatting and organization
+
+<a id="dates-times"></a>
+
+### Dates and times
+
+*Source: <https://developers.google.com/style/dates-times>*
+
+Expressing dates and times in a clear and unambiguous way helps support
+[writing for a global audience](#translation) and reduces
+confusion.
+
+<a id="dates-times--expressing-times"></a>
+
+#### Express times
+
+In general, use the following guidelines to format expressions of time:
+
+- Use the 12-hour clock, except if required to use a 24-hour time, such as
+  when documenting features that use 24-hour time. If the UI, a command, or a code sample uses the
+  24-hour format, use that format throughout the page for consistency.
+- Use exact times when possible, but *noon* and *midnight* are OK.
+- Use hyphens in time ranges. Don't add spaces before or after the hyphens.
+
+  Recommended: 5-10 minutes ago.
+- Capitalize AM and PM, and leave one space between it and the time.
+
+  Recommended: 3:45 PM.
+- Remove the minutes from round hours.
+
+  Recommended: 3 PM.
+
+<a id="dates-times--expressing-time-zones"></a>
+
+##### Express time zones
+
+Avoid using time zones unless absolutely necessary. In cases where you need to use a time
+zone—such as describing real events at real times—use the following guidelines:
+
+- Let the reader know if the time is local to their time—for example, *10 AM your local
+  time*.
+- If a time zone is necessary, use the timestamp format as seen in the user interface (if
+  available).
+- If using a specific time zone, spell out the region and include the
+  [UTC or GMT label](https://www.worldtimeserver.com/learn/utc-vs-gmt/)
+  as a parenthetical. For example:
+  - US and Canadian Pacific Standard Time (UTC-8)
+  - US and Canadian Pacific Daylight Time (UTC-7)
+- Don't abbreviate the name of the time zone.
+- In the rare event where the time of an event doesn't change for daylight saving time, use the
+  specific time zone, without reference to UTC.
+
+<a id="dates-times--expressing-dates"></a>
+
+#### Express dates
+
+In general, spell out the names of months and days of the week in full. Give
+the full four-digit year, not a two-digit abbreviation.
+
+Recommended: January 19, 2017
+
+If including the day of the week, add it before the month as follows:
+`DAY_OF_WEEK`, `MONTH` `DAY`,
+`YEAR`.
+
+Recommended: Tuesday, April 27, 2021
+
+<a id="dates-times--partial-dates-abbreviations"></a>
+
+##### Partial dates and abbreviations
+
+When giving only the month and year, don't use a comma.
+
+Recommended: She was hired in January
+2017.
+
+In most cases, don't abbreviate the day of the week or the month. However,
+when conserving space, such as in a heading or table, it's okay to abbreviate
+the month and the day of the week to their three-letter abbreviations.
+Capitalize the first letter and do not add a period at the end of the abbreviation.
+
+If you abbreviate, do so for the entire date. Don't mix written-out forms with
+abbreviated forms in the same date.
+
+Be consistent in where you apply abbreviations throughout your documentation. For
+example, if you choose to abbreviate in table cells, do so in all table cells.
+
+Recommended: Mon, Sep 3, 2018
+
+Not recommended: Mon, September 3, 2018
+
+<a id="dates-times--dates-mid-sentence"></a>
+
+##### Dates in the middle of a sentence
+
+When a `MONTH` `DAY`, `YEAR`
+date appears in the middle of a sentence, add a comma after the year.
+
+Recommended: The January 19, 2017,
+release of ...
+
+However, if the date in the middle of the sentence consists of the
+month and year only, don't use a comma.
+
+Recommended: The January 2017 release
+of ...
+
+<a id="dates-times--why-dates-written-out"></a>
+
+##### Why we prefer dates written out
+
+In general, don't express months as numbers unless you don't have the option
+(in which case, see [numeric-only date
+format](#dates-times--numeric-only-date-format)). Different regions of the world put parts of the date in a different
+order for numeric dates. For example, a date written as 04/05/09 means different
+things in different regions:
+
+- In the UK, 04/05/09 means May 4, 2009, where the order is usually day,
+  month, and then year.
+- In the US, 04/05/09 means April 5, 2009, where the order is usually month,
+  day, and then year.
+- In some other parts of the world, 04/05/09 means May 9, 2004. Some
+  regions write the year first, followed by the month and day.
+
+For this reason, we recommend always using words to express dates. Expressing
+dates in numbers only (using slashes, periods, or hyphens as separators) can be
+confusing.
+
+Recommended: February 12, 2017
+
+Recommended: Sunday, February 12, 2017
+
+Not recommended: 02.12.2017
+
+Not recommended: 12/02/2017
+
+<a id="dates-times--numeric-only-date-format"></a>
+
+##### Numeric-only date format
+
+If you must express a date in numerical date format, use the format
+`YYYY-MM-DD`, and separate the elements by using hyphens. This conforms
+to [ISO 8601 international
+standards](https://wikipedia.org/wiki/ISO_8601) for numerical date format.
+
+Additionally, if you have a choice of what date to write (such as in a
+fictional example), then choose a calendar day greater than 12 to differentiate
+it from the month.
+
+Recommended: 2017-04-15
+
+Not recommended: 04/06/2017
+
+<a id="dates-times--express-dates-and-times-together"></a>
+
+##### Express dates and times together
+
+If you must express a date and a time together, then mention the date first and then the time.
+
+Recommended: 2017-04-15 at 3 PM
+
+Recommended: May 4, 2009, at 6 PM
+
+<a id="dates-times--divisions-year"></a>
+
+#### Express divisions of the year
+
+Avoid referring to seasons. Spring in the northern hemisphere is fall (autumn) in the
+southern hemisphere. Instead, use the month, quarter, or temperature (if relevant).
+
+| Recommended | Not recommended |
+| --- | --- |
+| During warmer months, data centers face a higher risk of cooling failures. | During summer months, data centers face a higher risk of cooling failures. |
+| In November and December, data centers experience higher traffic volume. | In winter, data centers experience higher traffic volume. |
+| Changes are released in October of each year. | Changes are released in the Fall of each year. |
+
+---
+
+<a id="format-examples"></a>
+
+### Examples
+
+*Source: <https://developers.google.com/style/format-examples>*
+
+To introduce an example in a sentence, use the guidance in the following table. You can introduce
+examples using *such as*, *for example*, or *like* in various ways.
+
+| Guidance | Recommended | Not recommended |
+| --- | --- | --- |
+| Introduce a short-to-medium-length example at the end of a sentence. For clarity, consider setting off the example using a comma, parentheses, or an em dash as appropriate. Avoid using a semicolon for this purpose. | Choose a strong encryption algorithm, such as AES-256.  You can monitor various metrics for your managed database instances—for example, CPU utilization, storage capacity, and active connections.  The API supports common image formats like PNG and JPEG. | Enter a name for the instance, for example, `my-instance-99`.  Specify the region for deployment; for example, `us-central1`.  Enter a name for the instance (for example, `my-instance-99`). |
+| Introduce a short example in the middle of a sentence. Keep the example in the middle of a sentence relatively short and consider setting it off with dashes, commas, or parentheses as appropriate. | Enter a six-digit hex number (for example, `228B22`), and then click **OK**.  The virtual machine (VM) requires an operating system, such as Ubuntu 22.04, to be installed.  Some elements, like buttons and input fields, have default accessibility attributes. | Enter a six-digit hex number (for example, if you want the color forest green, enter `228B22`), and then click **OK**. |
+| Introduce a longer example as a separate sentence. For a longer example, introduce it as a separate sentence using *for example* as an adverb in that sentence. | You can assign tags to your virtual machine instances to categorize them. For example, you could tag instances by environment with `env:prod` or `env:dev`. |  |
+
+---
+
+<a id="images"></a>
+
+### Figures and other images
+
+*Source: <https://developers.google.com/style/images>*
+
+Use images only when they provide useful visual explanations of information
+that is otherwise difficult to express with words. For screenshots, be discreet. Only capture UIs
+that are important to the discussion.
+
+<a id="images--creating"></a>
+
+<a id="images--create-and-save-images"></a>
+
+#### Create and save images
+
+Consider the following guidelines for images:
+
+- To create a diagram, use any drawing tool.
+- To take a screenshot, use any screen capture tool.
+- Don't use images of text, code samples, or terminal output. Use actual
+  text.
+- For diagrams (architectural drawings, flow diagrams, and so on, as
+  distinct from screenshots), use the following guidelines:
+
+- Use SVG files if possible because SVGs stay sharp when you zoom in on
+  the image.
+- If you don't have an SVG file, then
+  save your image as a PNG file unless you have a good reason to use a
+  different format.
+- Regardless of the format, don't use a transparent background. In
+  particular, a transparent background can cause issues if you use the
+  Devsite lightbox widget.
+
+- For animations and videos, don't use animated GIF. Instead, use a more resource-efficient
+  format (such as MP4).
+- Be consistent for a given document or doc set in what operating system you use for
+  screenshots—for example, take all screenshots on macOS or on Linux. Similarly, be consistent
+  in how your screenshots look. If you take screenshots that include drop shadows of the
+  main window, make sure that similar screenshots are consistent.
+- Crop screenshots to show the relevant information. For example, don't include the
+  full window if you just want to show a single button or menu item. Cropping helps the
+  reader focus on the information that you want to convey in the screenshot, and it can help
+  future-proof the screenshot if other parts of the UI change.
+- Don't include personally identifying information (PII) in
+  screenshots.
+
+  If a source screenshot includes PII, hide it with a solid-color overlay
+  with 100% opacity. Don't rely on blurs, mosaic effects, or similar
+  image-processing effects to obscure PII; such effects can be reversed to reveal
+  the original information.
+
+  If you're exporting an image to a format that can include information on
+  separate layers (for example, PDF or TIFF), flatten the image on export.
+- Don't use image maps. Instead, provide a list of text references following the image. Reasons
+  to avoid image maps include the following:
+
+- Image maps are problematic for accessibility.
+- Browser implementation for image maps varies, and image maps might not function correctly on
+  mobile devices due to scaling.
+- The technical complexity of creating and maintaining a coordinates overlay is often
+  prohibitive.
+- Use descriptive filenames for your image files. For more information, see
+  [Filenames and file types](#filenames).
+
+<a id="images--text-with-images"></a>
+
+<a id="images--text-associated-with-images"></a>
+
+#### Text associated with images
+
+There are differences between alt text, figure captions, and figure
+descriptions. Independently of these elements, an introductory sentence should precede most images.
+The sentence can end with a colon or a period; usually a colon if it immediately precedes the image,
+usually a period if there's more material (such as a note paragraph) between the introduction and
+the image. Always introduce an image with a complete sentence. You don't need to introduce
+screenshots that immediately follow procedural text that describes a UI.
+
+<a id="images--example"></a>
+
+##### Example
+
+The following diagram shows how you can apply bounded contexts to an existing
+ecommerce application:
+
+<a id="images--bounded"></a>
+![Bounded contexts are applied to an application.](https://cloud.google.com/architecture/images/microservices-architecture-refactoring-monoliths-bounded-contexts.svg)
+
+
+**Figure 1**. Application capabilities are separated into bounded contexts that
+migrate to services.
+
+<a id="images--descr-1"></a>
+
+In figure 1, the ecommerce application's capabilities are separated into
+bounded contexts and migrated to services as follows:
+
+- Order management and fulfillment capabilities are bound into the
+  following categories:
+  - The order management capability migrates to the order service.
+  - The logistics delivery management capability migrates to the
+    delivery service.
+  - The inventory capability migrates to the inventory service.
+- Accounting capabilities are bound into a single category:
+  - The consumer, sellers, and third-party capabilities are bound
+    together and migrate to the account service.
+
+<a id="images--html"></a>
+
+##### HTML
+
+```
+      <p>The following diagram shows how you can apply bounded contexts to an existing ecommerce application:</p>
+<figure id="bounded">
+  <img src="https://cloud.google.com/architecture/images/microservices-architecture-refactoring-monoliths-bounded-contexts.svg"
+    alt="Bounded contexts are applied to an application.">
+  <figcaption><b>Figure 1.</b> Application capabilities are separated into bounded
+  contexts that migrate to services.</figcaption>
+</figure>
+<div id="descr-1">
+<p>In figure 1, the ecommerce application's capabilities are separated into
+bounded contexts and migrated to services as follows:</p>
+<ul>
+  <li>Order management and fulfillment capabilities are bound into the
+    following categories:
+  <ul>
+    <li>The order management capability migrates to the order service.</li>
+    <li>The logistics delivery management capability migrates to the
+        delivery service.</li>
+    <li>The inventory capability migrates to the inventory service.</li>
+  </ul>
+  </li>
+  <li>Accounting capabilities are bound into a single category:
+  <ul>
+  <li>The consumer, sellers, and third-party capabilities are bound
+        together and migrate to the account service.</li>
+  </ul>
+  </li>
+  </ul>
+</div>
+```
+
+<a id="images--markdown"></a>
+
+##### Markdown
+
+```
+The following diagram shows how you can apply bounded contexts to an existing ecommerce application:
+
+![Bounded contexts are applied to an application.](https://cloud.google.com/architecture/images/microservices-architecture-refactoring-monoliths-bounded-contexts.svg)
+
+**Figure 1.** Application capabilities are separated into bounded contexts that migrate to services.
+
+In figure 1, the ecommerce application's capabilities are separated into bounded contexts and
+migrated to services as follows:
+
+-   Order management and fulfillment capabilities are bound into the following categories:
+
+    -   The order management capability migrates to the order service.
+    -   The logistics delivery management capability migrates to the delivery service.
+    -   The inventory capability migrates to the inventory service.
+
+-   Accounting capabilities are bound into a single category:
+
+    -   The consumer, sellers, and third-party capabilities are bound together and migrate to the
+        account service.
+```
+
+<a id="images--alt-text"></a>
+
+##### Alt text
+
+*Alt text* is a concise description of the image that can replace the image in
+situations where the image isn't visible, such as people using screen
+readers, people using text-only browsers, or people who have a low-bandwidth
+internet connection. Alt text should consider the context of the image, not just its content.
+The presence of `alt` attributes helps support
+[navigability](https://web.dev/labels-and-text-alternatives/#include-text-alternatives-for-images-and-objects)
+in screen readers, [markup validation](https://validator.w3.org/docs/help.html#validation_basics),
+and [search engine optimization](https://support.google.com/webmasters/answer/7451184#usealtattribute).
+For more information, see [alt attribute](https://wikipedia.org/wiki/Alt_attribute).
+
+However, if the image is decorative (not informative) or it's
+provided only as a visual aid for information that is already expressed in text,
+then provide empty alternative text (`alt=""`) so it's ignored
+by assistive technologies. Examples of decorative images include the following:
+
+- A screenshot of the UI showing a user how to fill out fields.
+- Icons in the UI.
+- Images whose purpose is to make the page more visually appealing.
+
+When using the `img` element, the `alt` attribute is required,
+even if its assigned value is an empty string (`alt=""`). If you exclude the `alt`
+attribute completely, then screen readers might instead read the filename aloud.
+
+As per the [HTML
+specification](https://html.spec.whatwg.org/dev/images.html#general-guidelines), "the most general rule to consider when writing alternative
+text is the following: the intent is that replacing every image with the text of
+its `alt` attribute does not change the meaning of the page." So if the
+alternative text is redundant with surrounding text or it's not useful to
+visually impaired readers, use the empty tag.
+
+Consider the following when writing alt text:
+
+- Don't include phrases like *Image of* or *Photo of*.
+- Include punctuation. When screen readers encounter punctuation, they
+  pause before continuing.
+- Use consistent alt text for repeated instances of an image, such as
+  controls, status indicators, or icons that appear multiple times in your
+  document.
+- When possible, avoid using all-caps in alt text. Some screen readers read
+  capital letters as each letter individually.
+- Introduce diagrams in the text, not in the alt text.
+- Don't use figure captions to replace alt text.
+- Use full sentences or a noun phrase.
+
+  Recommended: `alt="Architecture of
+  an app that's built with Apps Script."`
+
+  Recommended: `alt="A card
+  message."`
+- Write short, descriptive alt text in 155 characters or less.
+- If the image presents more useful information than you can fit in the 155 character limit,
+  include a brief summary of the image in the `alt`attribute and also include a more
+  extensive description of the image in the text.
+- Alt text should consider the context of the image, not just its content.
+
+<a id="images--figure-captions"></a>
+
+##### Figure captions
+
+*Figure captions* are concise and comprehensive summaries of a figure or image. Figure
+captions (and figure numbers) are optional. When using the [`figcaption`
+element](https://html.spec.whatwg.org/multipage/semantics.html#the-figcaption-element), you must wrap both the `figcaption` and `img`
+elements in the [`figure`
+element](https://html.spec.whatwg.org/multipage/semantics.html#the-figure-element) to ensure that the figure caption is properly associated with the image.
+
+Consider the
+following when writing figure captions:
+
+- Figure numbers are optional. If you use figure numbers, use the form "<b>Figure
+  NUMBER.</b> DESCRIPTION."
+
+  Recommended: **Figure 1**. Application
+  capabilities are separated into bounded contexts that migrate to services.
+
+  Recommended: Application
+  capabilities are separated into bounded contexts that migrate to services.
+
+  Not recommended: Bounded contexts
+- We recommend using complete sentences in figure captions.
+- Always use end punctuation for captions.
+- When you refer to a figure, don't use spatial descriptions such as "the image above."
+
+  - If you used figure numbers, consistently refer to the figure by number. For example:
+    "... as shown in figure 1." Don't capitalize the word *figure* in a reference to a figure,
+    except at the start of a sentence.
+  - If you can't use figure numbers, show the figure again, for accessibility and user experience
+    reasons.
+- Don't include the figure caption in a sentence referencing the
+  figure.
+
+<a id="images--figure-descriptions"></a>
+
+##### Figure descriptions
+
+A *figure description* is text that provides a more detailed explanation of information
+represented by a figure. In other words, the information that is conveyed in the image is captured
+in the text. Any new information should be conveyed through text and not introduced in
+a figure or image.
+
+Consider the following when writing figure descriptions:
+
+- Create text that conveys the same information as the figure.
+- Use when a figure caption doesn't convey the purpose or complete information of the figure.
+- Use punctuation in figure descriptions.
+
+<a id="images--text-in-figures"></a>
+
+##### Text in figures
+
+In most cases, avoid embedding explanatory text in screenshot graphics; text
+that's incorporated into a graphic hurts accessibility and searchability, and
+increases localization costs if figures are localized. If you must embed text in
+an image, then be sure to also provide the same information in a form that
+people with visual disabilities can use, such as a figure description.
+
+When you must include text in figures and images, use the following
+guidelines:
+
+- Keep text brief. Avoid complete sentences and punctuation when
+  possible.
+- Don't embed figure descriptions or captions in the figure or image.
+  Instead, put figure descriptions and captions in text following the figure.
+- Don't create new abbreviations to condense text.
+- Use sentence case. Follow guidelines for [capitalization
+  for titles and headings](https://developers.google.com/style/capitalization#capitalization-in-titles-and-headings_1).
+- Use numbered callouts in figures to help you write a figure description,
+  but don't use callouts for detailed annotations in the image.
+- Use full trademarked product names.
+
+<a id="images--accessibility-resources"></a>
+
+##### Accessibility resources
+
+For more information about the accessibility of diagrams and screenshots, see the following
+resources:
+
+- [Web Content Accessibility Guidelines (WCAG)](https://www.w3.org/WAI/standards-guidelines/wcag/glance/)
+- [General text alternative guidelines from WCAG](https://www.w3.org/WAI/WCAG21/quickref/?showtechniques=111#text-alternatives)
+- [Using `alt` attributes for `img` elements](https://www.w3.org/WAI/WCAG21/Techniques/html/H37.html)
+- [Providing a long description in text near the non-text content](https://www.w3.org/WAI/WCAG21/Techniques/general/G74.html)
+- [Complex images](https://www.w3.org/WAI/tutorials/images/complex/)
+
+<a id="images--high-resolution"></a>
+
+<a id="images--high-resolution-images"></a>
+
+#### High-resolution images
+
+Modern browsers can use high-resolution images if they are available; this
+makes the images look better on high-resolution displays.
+
+To provide a high-resolution image, use the `img`
+element's `srcset` attribute in addition to the standard
+`src` attribute. The `srcset` attribute lets you specify
+different image assets for different screen resolutions. It accepts a
+comma-delimited set of image URLs, with the target screen resolution specified
+by a size qualifier: `1x` meaning the "standard" resolution,
+`2x` meaning "double" the resolution, and so on.
+
+If a web browser supports the `srcset` attribute, it selects an
+image from the specified images that's an appropriate resolution for the current
+display. If the browser doesn't support the `srcset` attribute, it
+uses the image in the `src` attribute. Consequently, you must always
+still include the `src` attribute.
+
+For example, to provide both a standard resolution image and a
+double-resolution image, add a `srcset` attribute and specify both
+`1x` and `2x` image assets:
+
+```
+<img src="/assets/images/skateboard.png"
+  srcset="/assets/images/skateboard.png 1x,
+  /assets/images/skateboard_2x.png 2x"
+  width="375" alt="" />
+```
+
+- The `width` attribute matches the CSS pixel size used for the
+  page dimensions. (The height is automatically calculated based on the width and
+  the image's proportions; *don't* state it explicitly.)
+- Set the `src` attribute to point to the standard-resolution
+  (`1x`) image, *not* the `2x` version. (Almost
+  everyone who has a high-resolution screen also has a modern browser that can
+  recognize the `srcset` attribute. The `src` attribute is
+  mainly used by older browsers on low-resolution devices, which should download
+  the smaller, low-resolution image.) Even if your original image is the
+  higher-resolution image, set the `src` attribute to use the
+  standard-resolution version; don't force a reader using a low-resolution screen
+  to download a graphic that's higher-resolution than they can view.
+- The filename for the double-resolution image (in this case,
+  `skateboard_2x.png`) can be anything—it's the "`2x`"
+  value following the filename that informs the browser which resolution the file
+  is. But it's a good idea to use a filename of the form
+  `BASENAME_2x.EXTENSION` to make clear to human
+  readers that it's a double-resolution version of
+  `BASENAME.EXTENSION`.
+- The double-resolution image must be exactly twice the width and height of
+  the standard image, give or take a pixel. (For example, it's okay for the
+  double-resolution image to be 875x500 and the standard size to be 438x250.)
+- Don't scale up an existing `1x` image to make the
+  `2x` version. If all you have is the `1x` version, then
+  use it alone. But if you're starting with a high-resolution image (at
+  `2x` resolution or better), then you can scale it down to appropriate
+  dimensions for `1x` and `2x`.
+- Currently, only an additional `2x` image is necessary, but
+  someday screen PPI may increase further.
+  So the `srcset` attribute supports further alternative sizes, each
+  specified by the appropriate multiplier, such as `3x` or
+  `4x`.
+- A browser that supports the `srcset` attribute uses only the
+  images provided in that attribute—it ignores the `src` attribute. So
+  specify all available image resolutions in the `srcset`
+  attribute.
+
+> [!NOTE]
+> **Note**: If you frequently revise an image, then you can
+> use the `2x` image for both the `src` and
+> `srcset` attributes, rather than maintaining multiple sizes of the
+> image. If things stabilize and you no longer need to revise the image, then you
+> can add a `1x` version.
+
+For more information, see the HTML specification for the `img`
+element.
+
+<a id="images--layout"></a>
+
+<a id="images--layout-of-images-on-a-page"></a>
+
+#### Layout of images on a page
+
+Consider the following guidelines for adding images to pages:
+
+- Don't try to place an image manually; for example, don't use a
+  `style` attribute or other workarounds to control the image's
+  left/right justification or the margins around the image. Instead, use
+  your site's standard CSS image styles.
+- Don't make your image too small. It's fine for an image to take up the
+  full width of a page.
+- Consider how the image will look when printed out.
+- In general, don't use an image that's wider than the column it appears
+  in. On [developer.android.com](https://d.android.com), for
+  example, the main-body column is 856px wide, so use images that are no wider
+  than that. In that context, the high-resolution 2x version of the image should
+  be no wider than 1712px.
+
+  - Screenshots at full resolution often take up too much space on the
+    page, so you may have to resize them.
+  - If the graphics were created by someone else (for example, a designer
+    on the team you're supporting), it may be fairly trivial for them to provide you
+    with images at the appropriate size. If the images they provide are wider than
+    856px, ask the designer if they can provide the relevant graphics as
+    856px/1712px pairs.
+- Don't link to the figure from within the same page unless it's a very
+  long page and you're linking to it from quite far away on the page.
+- Don't center the image on the page.
+- Don't put an `img` element inside a `p` element.
+
+---
+
+<a id="footnotes"></a>
+
+### Footnotes
+
+*Source: <https://developers.google.com/style/footnotes>*
+
+A footnote is an annotation with additional information usually provided at the end of a page,
+chapter, or book. We recommend avoiding footnotes because they aren't accessible and can present
+challenges for localization efforts.
+
+Instead of a footnote, consider using the following formats to convey information:
+
+- [Add a cross-reference](#cross-references).
+- [Use a note](#notices).
+- [Put it in a parenthetical](#parentheses).
+
+If the only way to convey this information is to use a footnote, then use a superscript
+number—for example, `<sup>1</sup>`.
+
+Recommended: You want to add a footnote to this sentence.1
+
+1 Put this footnote at the bottom of the page.
+
+---
+
+<a id="headings"></a>
+
+### Headings and titles
+
+*Source: <https://developers.google.com/style/headings>*
+
+Use sentence case for headings and titles. Use descriptive headings and titles because they help
+a reader navigate their browser and the page. It's easier to jump between pages and sections of a
+page if the headings and titles are unique.
+
+<a id="headings--heading-and-title-text"></a>
+
+#### Heading and title text
+
+Write document titles based on the primary purpose of the document. If a
+document is primarily a tutorial, but it has a conceptual introduction, write a
+task-based title. Write section headings based on the type of content that's in
+the section.
+
+| Guidance | Recommended | Not recommended |
+| --- | --- | --- |
+| For a task-based heading, start with a [*bare infinitive*](https://wikipedia.org/wiki/Infinitive#English), also known as a *plain form* or [*base form*](https://wikipedia.org/wiki/English_verbs#Base_form) verb. In English, the *imperative mood* also uses the base form verb, so it looks the same as the bare infinitive.  Task-based headings are frequently used in quickstarts, how-to documents, and tutorials. | Create an instance | Creating an instance |
+| For a conceptual or non-task-based heading, use a [*noun phrase*](https://wikipedia.org/wiki/Noun_phrase) that doesn't start with an *-ing* verb.  Noun-phrase headings are frequently used in concept documentation. | Migration to Google Cloud | Migrating to Google Cloud |
+| If a section is not required for all users or scenarios, use the *Optional:* prefix in the heading. This prefix signals when the section information applies only to a specific configuration or use case.  For information about optional steps in a procedure, see [Optional steps](#procedures--optional-steps). | Optional: Customize your alias | Customize your alias (optional) |
+
+<a id="headings--title-phrasing"></a>
+
+##### Title phrasing
+
+Use a unique level-1 heading (`h1`) for each page in a set of documents and only use
+a level-1 heading once on a page.
+
+Avoid repeating the exact page title in a heading on the page. For example, if you document
+how to create a virtual machine and how to start a virtual machine on the same task-based
+page, the page title might be *Create and start VM instances*, with section headings
+*Create a VM* and *Start a VM*.
+
+<a id="headings--mixed-heading-styles"></a>
+
+##### Mixed heading styles
+
+It's OK to use task-based and conceptual heading styles in the same document.
+If a single document includes both task-based and conceptual sections, then use
+the appropriate phrasing for each section's heading.
+
+<a id="headings--use-of--ing-verb-forms"></a>
+
+##### Use of *-ing* verb forms
+
+When possible, avoid using *-ing* verb forms as the first word in any heading or
+title.
+
+Recommended:
+Transfer data sets
+
+Not recommended:
+Transferring data sets
+
+An *-ing* verb form is a present participle or gerund. These verb forms
+are inconsistently translated when they're used as the first word in a title,
+and they increase character count in limited spaces.
+
+Sometimes, there might not be a better alternative to using a gerund, such as the following
+examples:
+
+- Billing
+- Pricing
+
+It's OK to use a gerund in these cases.
+
+It's OK to use an *-ing* verb form later in a heading or title, such as
+*Introduction to BigQuery monitoring*.
+
+<a id="headings--example-headings"></a>
+
+##### Example headings
+
+The following example is a task-based document that includes a conceptual
+heading and a task-based heading.
+
+Recommended:
+
+<a id="headings--html"></a>
+
+##### HTML
+
+```
+<h1>Log serving requests by using AI Platform Prediction</h1>
+
+<p>This task-based document shows how to monitor machine learning models. The
+document title starts with a bare infinitive.</p>
+
+<h2>ML model monitoring overview</h2>
+
+<p>This section provides a conceptual overview of ML model monitoring. Its title is
+a noun phrase.</p>
+
+<h2>Configure notebook settings<h2>
+
+<p>This task-based section provides a series of steps to set variables in a
+notebook. Its title starts with a bare infinitive.</p>
+```
+
+<a id="headings--markdown"></a>
+
+##### Markdown
+
+```
+# Log serving requests by using AI Platform Prediction
+
+This task-based document shows how to monitor machine learning models. The
+document title starts with a bare infinitive.
+
+## ML model monitoring overview
+
+This section provides a conceptual overview of ML model monitoring. Its title is
+a noun phrase.
+
+## Configure notebook settings
+
+This task-based section provides a series of steps to set variables in a
+notebook. Its title starts with a bare infinitive.
+
+```
+
+<a id="headings--capitalization"></a>
+
+<a id="headings--heading-and-title-format"></a>
+
+#### Heading and title format
+
+The following sections list and define our writing standards for capitalization,
+abbreviations, and technical elements in headings. In general, guidance
+that applies to standard text also applies to headings—for example,
+[contractions](#contractions) and [articles](#articles).
+
+<a id="headings--syntax-and-capitalization"></a>
+
+##### Syntax and capitalization
+
+- **Use sentence case** for all headings and titles. For more information, see
+  [Capitalization in titles and
+  headings](#capitalization--capitalization-in-titles-and-headings).
+- **Keep punctuation simple**. Punctuation can be a sign that your heading is too complicated. Consider rewriting.
+- **Limit abbreviations**. Only use an abbreviation of a word in a page title or heading if it's the more commonly known
+  version of the word. If you do so, define the abbreviation in the first instance of the word in a paragraph.
+  You can define the abbreviation in the page title or heading, but consider if the additional
+  length adds value. For SEO, use the more prominent version of a term in headings. For more information, see
+  [Abbreviations](#abbreviations).
+
+<a id="headings--formatting-and-code"></a>
+
+##### Formatting and code
+
+- **Don't use numbers in headings** to indicate a sequence of sections.
+  Instead, rely on heading hierarchy and order to indicate sequence.
+- **Avoid code items in headings**. If you must mention a code item in a heading,
+  add a descriptive noun to the item in code font. For more information, see
+  [Grammatical treatment of code elements](#code-in-text--grammatical-treatment-of-code-elements).
+- **Don't put links in headings**. A link can easily be confused as a style applied to the
+  heading instead of a link.
+
+<a id="headings--hierarchy-and-structure"></a>
+
+##### Hierarchy and structure
+
+- **Don't use heading tags to change visual formatting**. Use CSS rather than a heading level
+  that doesn't fit the hierarchy. Don't make up your own formatting for headings.
+- **Apply proper heading tags**. Use heading tags to structure your content hierarchically—for example,
+  `<h1>`, `<h2>`, and `<h3>` in HTML, or
+  `#`, `##`, and `###` in Markdown.
+- **Maintain logical order**. Don't skip levels of the heading hierarchy. For example, put an `<h3>` tag
+  only under an `<h2>` tag.
+
+  Recommended:
+
+  <a id="headings--html_1"></a>
+
+  ##### HTML
+
+  ```
+  <h1>Transfer data sets</h1>
+
+  <p>This document provides a high-level overview of ways to transfer your data to Google
+  Cloud.</p>
+
+  <h2>Estimate costs</h2>
+  ```
+
+  <a id="headings--markdown_1"></a>
+
+  ##### Markdown
+
+  ```
+  # Transfer data sets
+
+  This document provides a high-level overview of ways to transfer your data to Google Cloud.
+
+  ## Estimate costs
+  ```
+
+  Not recommended:
+
+  <a id="headings--html_2"></a>
+
+  ##### HTML
+
+  ```
+  <h1>Transfer data sets</h1>
+
+  <p>This document provides a high-level overview of ways to transfer your data to Google
+  Cloud.</p>
+
+  <h3>Estimate costs</h3>
+  ```
+
+  <a id="headings--markdown_2"></a>
+
+  ##### Markdown
+
+  ```
+  # Transfer data sets
+
+  This document provides a high-level overview of ways to transfer your data to Google Cloud.
+
+  ### Estimate costs
+  ```
+- **Don't use empty headings**. Make sure headings are followed by content.
+
+  Recommended:
+
+  <a id="headings--html_3"></a>
+
+  ##### HTML
+
+  ```
+  <h2>Migrate VMs to Compute Engine</h2>
+
+  <p>Migration is not just a single step. The following sections describe the recommended
+  steps.</p>
+
+  <h3>Design the migration</h3>
+  ```
+
+  <a id="headings--markdown_3"></a>
+
+  ##### Markdown
+
+  ```
+  ## Migrate VMs to Compute Engine
+
+  Migration is not just a single step. The following sections describe the recommended steps.
+
+  ### Design the migration
+  ```
+
+  Not recommended:
+
+  <a id="headings--html_4"></a>
+
+  ##### HTML
+
+  ```
+  <h2>Migrate VMs to Compute Engine</h2>
+
+  <h3>Design the migration</h3>
+  ```
+
+  <a id="headings--markdown_4"></a>
+
+  ##### Markdown
+
+  ```
+  ## Migrate VMs to Compute Engine
+
+  ### Design the migration
+  ```
+
+<a id="headings--introducing-multiple-sections"></a>
+
+<a id="headings--refer-to-a-group-of-sections"></a>
+
+#### Refer to a group of sections
+
+If you introduce a group of related H3 or lower sections within a larger H2 section, use the
+phrase *the following sections*. Don't refer to the group of sections using the phrases
+*this section* or *these sections* because those phrases are ambiguous.
+
+Recommended:
+
+<a id="headings--html_5"></a>
+
+##### HTML
+
+```
+<h2>Views in the data preparation editor</h2>
+
+<p>The following sections describe the views in the data preparation editor.</p>
+
+<h3>Data view</h3>
+
+<p>...</p>
+
+<h3>Graph view</h3>
+
+<p>...</p>
+
+<h3>Schema view</h3>
+
+<p>...</p>
+```
+
+<a id="headings--markdown_5"></a>
+
+##### Markdown
+
+```
+## Views in the data preparation editor
+
+The following sections describe the views in the data preparation editor.
+
+### Data view
+
+...
+
+### Graph view
+
+...
+
+### Schema view
+
+...
+```
+
+---
+
+<a id="italics-terms"></a>
+
+### Italics with terms
+
+*Source: <https://developers.google.com/style/italics-terms>*
+
+This page describes two circumstances when we italicize terms that we're
+introducing or discussing.
+
+For more information about italics and other formatting, including HTML and
+Markdown formatting for italics, see
+[Text-formatting summary](#text-formatting).
+
+<a id="italics-terms--new_terms"></a>
+
+#### New terms
+
+When you introduce a new term that you're defining immediately, use italics on
+the first mention of the term. Don't use bold or quotation marks.
+
+Recommended: A
+*Clos network* is a kind of multistage circuit switching network.
+
+<a id="italics-terms--words_as_words"></a>
+
+#### Words as words
+
+When you refer to a word, phrase, or letter in reference to the word, phrase,
+or letter itself (sometimes called *words as words*) use italics. Don't use bold
+or quotation marks.
+
+Recommended: Don't use
+*&* (ampersand) as a conjunction. Use the word *and* instead.
+
+Recommended: To form a
+possessive of a singular noun, add *'s* to the end of the word.
+
+---
+
+<a id="lists"></a>
+
+### Lists
+
+*Source: <https://developers.google.com/style/lists>*
+
+<a id="lists--list-or-table"></a>
+
+#### List or table?
+
+Tables and lists are both ways to present a set of similarly structured
+items. Sometimes it's not obvious when to choose one presentation over the
+other. To decide which presentation to use, see
+[List or table?](#tables--list-or-table)
+
+**Note**: Don't use a list to show only one item; a single
+item isn't really a list. If you want to set a single item off from surrounding
+text, then use some other formatting.
+
+<a id="lists--types"></a>
+
+<a id="lists--types-of-lists"></a>
+
+#### Types of lists
+
+Choose one of the following list styles. The following table includes common ways to present
+lists in our documentation:
+
+| List type | Used for | HTML elements |
+| --- | --- | --- |
+| Numbered list | A set of items where the sequence is significant, such as ordered steps, phases, or priorities. The following is an example of a numbered list:  Here's a sequence of steps to follow:   1. Open the box. 2. Remove the bobcat from the box. 3. Feed the bobcat.  Nested sequential lists are labeled with lowercase letters or lowercase Roman numerals. The following is an example of a nested sequential list:  Here's a list of things to do after breakfast, in order:   1. Go shopping.    1. Buy groceries:       - Flour       - Eggs       - Sugar       - Butter    2. Go to mall:       1. Buy dress.       2. Buy shoes. 2. Make cake. 3. Build birthday present out of spare parts. 4. Clean house.  See also [Sub-steps in numbered procedures](#procedures--sublists). | `ol`, `li` |
+| Bulleted list | A set of items that's not a sequence, such as a set of nonsequential options or examples. Make sure it's clear whether or not every item is required. The following is an example of a bulleted list:  Here's a list of things that can go wrong, in no particular order:   - Your bicycle might explode. - The sun might go out. - An ant might break its leg and require a tiny splint. | `ul`, `li` |
+| Description list | A set of terms, each with a description, definition, or explanation. Use this type of list if you want to draw attention to two or more terms (such as a glossary). The following is an example of a description list:  Here are some descriptions of types of birds:  Emu  The best kind of bird.  Crow  The other best kind of bird.  Peacock  Also the best kind of bird.  Phoenix  An even better kind of bird. | `dl`, `dt`, `dd` |
+| Description list that uses bulleted run-in headings | A set of introductory terms or phrases, each followed by a description, definition, or explanation. Use this type of list if you want to highlight and explain several concepts or save space. For information about how to format and punctuate run-in headings and their descriptions, see [Description lists that use run-in headings](#lists--description-lists-that-use-run-in-headings) in this document.  The following is an example of a description list that uses bulleted run-in headings:  Here are some descriptions of types of birds:   - **Emu**: the best kind of bird - **Crow**: the other best kind of bird - **Peacock**: also the best kind of bird - **Phoenix**: an even better kind of bird | `ul`, `li` |
+
+<a id="lists--multiple-paragraphs"></a>
+
+<a id="lists--multiple-paragraph-list-items"></a>
+
+#### Multiple paragraph list items
+
+Any list item can contain more than one paragraph.
+
+To create multiple paragraphs, use the `p` element rather
+than using the `br` element. (The HTML specification describes which uses of the
+[`br`
+element](https://html.spec.whatwg.org/multipage/semantics.html#the-br-element) are legitimate and which aren't.)
+
+Example of a list item that contains more than one paragraph:
+
+- This list item is a single paragraph.
+- This list item contains multiple paragraphs.
+
+  As you can see!
+- This is another list item that's only one paragraph long.
+
+<a id="lists--intros"></a>
+
+<a id="lists--introductory-sentences-for-lists"></a>
+
+#### Introductory sentences for lists
+
+Introduce a list with the appropriate context. In most cases, precede a list
+with an introductory sentence. The sentence can end with a colon or a period; usually a colon if it
+immediately precedes the list, usually a period if there's more material (such as a note
+paragraph) between the introduction and the list.
+
+If the list doesn't need any additional context other than the heading that immediately precedes
+the list, it's OK to not introduce a list with an introductory sentence.
+
+Introduce a list with a complete sentence, not a partial one that's
+completed by the list items. You can also use *the following* as a noun phrase (see
+[following](#word-list--following) in the word list).
+
+| Recommended | Not recommended |
+| --- | --- |
+| Use the **Submit** button for any of the following purposes:   - To submit the form. - To indicate that you're done. - To allow the next person to enter their data. | Use the **Submit** button to:   - Submit the form. - Indicate that you're done. - Allow the next person to enter their data. |
+| To get the USB driver, follow these steps:   1. Click **Tools >    Android > SDK Manager**. 2. Select **Google USB Driver**, and then click **OK**. | To get the USB driver:   1. Click    **Tools > Android >    SDK Manager**. 2. Select **Google USB Driver**, and then click **OK**. |
+| If you need to add an instance manually, do the following:   1. Click **Create instance**. 2. For **Name**, enter a name. | If you need to add an instance manually:   1. Click **Create instance**. 2. For **Name**, enter a name. |
+| <a id="lists--objectives"></a>Objectives  - Create an instance - Snapshot an instance - Delete an instance | <a id="lists--objectives_1"></a>Objectives In the following tutorial, you will complete the following tasks:   - Create an instance - Snapshot an instance - Delete an instance |
+
+For information about introducing sub-steps, see [Sub-steps in numbered procedures](#procedures--sublists).
+
+For information about punctuation and capitalization of lists, see
+[Capitalization and end punctuation](#lists--capitalization).
+
+<a id="lists--numbering"></a>
+
+<a id="lists--unusual-list-numbering"></a>
+
+#### Unusual list numbering
+
+Use nonstandard numbering in the following situations:
+
+- To present a list in reverse-numerical order, use an
+  `ol` element with a `reversed` attribute.
+- To set a value manually, use the `value` attribute. In some cases, setting a
+  value manually can be convenient. However, in most cases, it isn't a good idea to manually
+  number a list item in a numbered list, because if the number of items changes later, you'll
+  have to manually change the value.
+
+<a id="lists--sublists"></a>
+
+<a id="lists--sub-steps-in-a-numbered-procedure"></a>
+
+#### Sub-steps in a numbered procedure
+
+For information about sub-steps in a numbered procedure, see
+[Procedures](#procedures--sublists).
+
+<a id="lists--parallel"></a>
+
+<a id="lists--parallel-syntax"></a>
+
+#### Parallel syntax
+
+Use the same syntax/structure for all list items in a given list, if
+possible.
+
+<a id="lists--capitalization"></a>
+
+<a id="lists--capitalization-and-end-punctuation"></a>
+
+#### Capitalization and end punctuation
+
+Capitalization and end punctuation depend on the type of list and the
+contents of the list.
+
+<a id="lists--numbered-lettered-bulleted-lists"></a>
+
+##### Numbered, lettered, and bulleted lists
+
+Start each list item with a capital letter, unless case is an important part of
+the information conveyed by the list—such as in a list of glossary terms.
+
+End each list item with a period or other appropriate sentence-ending
+punctuation, except in the following cases:
+
+- If the item consists of a single word, don't add end punctuation.
+- If the item doesn't include a verb, don't add end punctuation.
+- If the item is entirely in code font, don't add end punctuation.
+- If the item is entirely link text or a document title, don't add end punctuation.
+
+If you end up with inconsistent punctuation in your list, then either rewrite your list to use
+[parallel construction](#lists--parallel) or add end punctuation to every list item
+for consistency.
+
+Recommended:
+
+The following words are adjectives:
+
+- Big
+- Small
+- Gratuitous
+
+Recommended:
+
+The SDK supports the following UI elements:
+
+- Text box
+- Bulleted list
+- Button
+
+Recommended:
+
+The API supports the following actions:
+
+- Create
+- Replace
+- Update
+- Delete
+
+Recommended:
+
+You can do any of the following by using the API:
+
+- Create an item.
+- Replace one item with another.
+- Update an item.
+- Delete an item.
+
+<a id="lists--description-lists"></a>
+
+##### Description lists
+
+Sometimes it's useful to add an explanatory phrase to a list item, which can
+affect the punctuation. In general, don't add an explanatory phrase to only a
+single list item; instead, use a description list, and provide explanatory
+phrases for all items.
+
+In most contexts, start each term (`dt` element) with a capital letter.
+
+Don't end the term with a period. Do generally put a period at the end of
+each `dd` ("description") element.
+
+| Recommended | Not recommended |
+| --- | --- |
+| The following words are adjectives:  Big  A short word.  Relevant  A fancy word.  Gratuitous  A long word.  Purple  A vibrant color. | The following words are adjectives:   - Big - Relevant - Gratuitous - Purple—this is a color. |
+
+<a id="lists--description-lists-that-use-run-in-headings"></a>
+
+##### Description lists that use run-in headings
+
+In most contexts, format run-in headings as follows:
+
+- Start the run-in heading with a capital letter.
+- End the run-in heading with a period or a colon, but be consistent within the list.
+- You can decide whether to bold the punctuation that ends the heading based on factors
+  such as on-page consistency.
+
+For the descriptions that follow the punctuation, capitalize the first letter as follows:
+
+- If the text follows a period, start the text with a capital letter.
+- If the text follows a colon, start the text with a lowercase letter.
+
+To end the descriptive text, punctuate as follows:
+
+- If the description follows a period, end the description with a period.
+- If the description follows a colon, do one of the following:
+
+- If the description is a list of items or short phrases without verbs, don't include a
+  period.
+- If the description includes a verb or expresses a standalone thought, end the
+  description with a period.
+
+Don't use a dash to set off a description from an item in a description list. For more
+information, see
+[Colons instead of dashes in
+lists](#dashes--colons-instead-of-dashes-in-description-lists).
+
+Recommended:
+
+The following words are adjectives:
+
+- **Big**: a short word
+- **Relevant**: a fancy word
+- **Gratuitous**: a long word
+- **Purple**: a vibrant color
+
+Recommended:
+
+The coffee shop has several great choices:
+
+- **Coffee**: latte, mocha, cappuccino, espresso, macchiato
+- **Tea**: chai tea, chai latte, black tea, green tea, herbal tea
+
+Recommended:
+
+Budget Airlines reduces your ticket cost in several ways:
+
+- **It increases fuel economy by reducing baggage weight**. By charging
+  astronomical prices for anything larger than a wallet....
+- **It carries more passengers per flight**. By reducing leg room to industry and
+  medical minimums, it fits more seats....
+
+> [!NOTE]
+> **Note**: The guidelines here about list punctuation
+> differ from the [Material
+> Design guidelines](https://material.io/guidelines/style/writing.html#writing-capitalization-punctuation). If you're writing UI text rather than prose
+> documentation, then follow the Material Design guidelines.
+
+<a id="lists--comma-separated-lists"></a>
+
+#### Comma-separated lists
+
+When you write a list in a paragraph, use
+[serial commas](#commas--serial-commas) to separate the items.
+
+Avoid ending a list with *etc.* or phrases like *and so on*.
+Instead, introduce the list in a way that makes it clear that the list isn't
+all-inclusive.
+
+Recommended: The service processes data
+like event logs, clickstream data, social network interactions, and e-commerce
+transactions.
+
+Not recommended: The service processes
+event logs, clickstream data, social network interactions, e-commerce
+transactions, etc.
+
+For more information, see [etc.](#word-list--etc)
+
+---
+
+<a id="mathematical-notation"></a>
+
+### Mathematical notation
+
+*Source: <https://developers.google.com/style/mathematical-notation>*
+
+This page describes how to format common mathematical notation such as
+exponents, expressions, equations, operators, and variables in
+documentation. Formatting best practices can help ensure that your
+documentation is compatible with assistive technologies and
+renders accurately.
+
+For general information about using and formatting numbers, see
+[Numbers](#numbers).
+
+**Note:** This page includes examples of formatting in HTML
+and Markdown in standard text. If you're using a third-party tool to display
+complex math, follow that tool's formatting guidance to ensure that your
+mathematical markup displays correct.
+
+<a id="mathematical-notation--html-entities"></a>
+
+<a id="mathematical-notation--use-html-entities-for-mathematical-symbols"></a>
+
+#### Use HTML entities for mathematical symbols
+
+In general, use HTML entities for mathematical symbols instead of keyboard
+symbols. The following table lists entities for symbols that are common in
+arithmetic and algebra. For the plus sign (`+`), equals sign
+(`=`), and division sign (`/`), you can use
+their keyboard equivalents.
+
+| Symbol | Markup | Description |
+| --- | --- | --- |
+| + | Use the keyboard symbol. | Plus sign |
+| − | `&minus;` | Minus sign |
+| × | `&times;` | Multiplication sign  Alternatively, you can use the dot operator `∙` (`&#8729;`) or asterisk operator `*` (`&#42;`) to match the UI. Don't use an asterisk (`*`) to indicate multiplication in text.  You can indicate multiplication by omitting the multiplication symbol if doing so doesn't create ambiguity—for example, instead of *a* × *b*, you can write *ab*. |
+| / | Use the keyboard symbol. | Division sign |
+| = | Use the keyboard symbol. | Equals sign |
+| ≠ | `&ne;` | Not equal to |
+| ± | `&plusmn;` | Plus-minus sign |
+| ∓ | `&mnplus;` | Minus-plus sign |
+| < | `&lt;` | Less than sign |
+| > | `&gt;` | Greater than sign |
+| ≈ | `&asymp;` | Approximately equal to |
+| ≉ | `&nap;` | Not approximately equal to |
+| ≅ | `&cong;` | Congruent to |
+| ≤ | `&le;` | Less than or equal to |
+| ≥ | `&ge;` | Greater than or equal to |
+| ≡ | `&equiv;` | Identical to |
+| ≢ | `&nequiv;` | Not identical to |
+| √ | `&radic;` | Square root |
+| ∑ | `&sum;` | N-ary summation |
+
+<a id="mathematical-notation--format-mathematical-notation"></a>
+
+#### Format mathematical notation
+
+The following sections provide formatting for common math-related notation.
+
+<a id="mathematical-notation--operators"></a>
+
+##### Operators
+
+To ensure accessibility and accurate HTML syntax, use
+[HTML entities](#mathematical-notation--html-entities)
+instead of keyboard symbols for operators. For example, use
+`&minus;` instead of a hyphen (`-`).
+
+Include a non-breaking space (`&nbsp;`) on both sides
+of operators within a single expression, equation, or
+statement.
+
+Don't italicize operators.
+
+Recommended: *a* − *b*
+
+To render *a* − *b*, use the following markup:
+
+- **HTML:** `<i>a</i>&nbsp;&minus;&nbsp;<i>b</i>`
+- **Markdown:** `_a_&nbsp;&minus;&nbsp;_b_`
+
+<a id="mathematical-notation--variables"></a>
+
+##### Variables
+
+Italicize variables.
+
+Recommended: *x* ≠ *y*
+
+Recommended: *x**y*
+
+Recommended: *y**i*
+
+To render *x* ≠ *y*, use the following markup:
+
+- **HTML:** `<i>x</i>&nbsp;&ne;&nbsp;<i>y</i>`
+- **Markdown:** `_x_&nbsp;&ne;&nbsp;_y_`
+
+<a id="mathematical-notation--expressions-equations"></a>
+
+##### Expressions and equations
+
+Include short expressions and equations inline with your text.
+
+Include a non-breaking space (`&nbsp;`) between components
+such as operators and variables so that the expression or equation renders on
+the same line.
+
+When an expression or equation creates an awkward line break, consider
+placing it on its own line.
+
+Recommended: The equation
+that describes a linear trend line is
+*y* = *a* + *bx*.
+
+Recommended: The equation
+that describes a polynomial trend line, where the order is *o*, is the
+following:
+*y* = *a* + *b* × *x* + ... + *k* × *x**o*
+
+To render *y* = *a* + *bx*, use the
+following markup:
+
+- **HTML:** `<i>y</i>&nbsp;&=&nbsp;<i>a</i>&nbsp;+&nbsp;<i>bx</i>`
+- **Markdown:** `_y_&nbsp;&=&nbsp;_a_&nbsp;+&nbsp;_bx_`
+
+<a id="mathematical-notation--fractions"></a>
+
+##### Fractions
+
+Express fractions as decimal numbers, when possible.
+
+If you must express fractions as words, connect the numerator and
+denominator with a hyphen unless one of them is already hyphenated.
+
+Recommended: 0.02
+
+Recommended: one and one-half
+
+Recommended: three-sevenths
+
+Recommended: three seventy-fourths
+
+<a id="mathematical-notation--exponents-subscripts"></a>
+
+##### Exponents and subscripts
+
+Use
+[standard mathematical notation](https://wikipedia.org/wiki/Exponentiation).
+Don't put a space between the base and the exponent.
+
+To render exponents, use the HTML `<sup>` tag. Don't use the
+keyboard caret symbol (`^`) to indicate an exponent.
+
+To render subscripts, use the HTML `<sub>` tag.
+
+Recommended: 23
+
+Recommended: *x**y*
+
+Recommended: *y**i*
+
+Not recommended: 2^3
+
+To render 23, use the following markup in HTML and Markdown:
+`2<sup>3</sup>`
+
+<a id="mathematical-notation--expressions-as-words"></a>
+
+<a id="mathematical-notation--notation-as-words"></a>
+
+#### Notation as words
+
+In general, you can use mathematical notation in place of words in running
+text. For example, in a sentence, you might use the statement
+*x* ≠ *y* instead of writing "*x* is not equal to
+*y*." If the use of notation instead of words creates ambiguous, grammatically
+incorrect, or difficult-to-read text, then use words to convey the
+mathematical concept.
+
+Recommended: Check
+whether *a* > *b*.
+
+Recommended: The area
+is calculated by multiplying the length by the width.
+
+Not recommended: Check
+whether *a* is greater than *b*.
+
+Not recommended: The
+area is calculated by multiplying *l* × *w*.
+
+<a id="mathematical-notation--math-concepts"></a>
+
+<a id="mathematical-notation--tools-for-complex-or-multiline-equations"></a>
+
+#### Tools for complex or multiline equations
+
+The methods described on this page that use HTML entities and tags are suitable for most common
+mathematical notation. However, for more complex, multiline equations, or formulas that are
+difficult to represent clearly with standard HTML, consider using [diagrams,
+other images](#images), or a dedicated math rendering tool to support comprehension. Images and diagrams
+like pie charts or bar graphs, in particular, are especially helpful for comparing statistics and
+illustrating percentages.
+
+<a id="mathematical-notation--more-resources"></a>
+
+#### More resources
+
+- [Numbers](#numbers)
+- [Units of measure](#units-of-measure)
+- [Text-formatting summary](#text-formatting)
+
+---
+
+<a id="notices"></a>
+
+### Notes and other notices
+
+*Source: <https://developers.google.com/style/notices>*
+
+To give the reader important or useful information that isn't part of the flow of the text, you
+can offset the information with a notice. However, there's
+[evidence](https://www.nngroup.com/articles/tunnel-vision-and-selective-attention/)
+that readers skip elements on the page, including notices, that are outside their focus of
+interest. If you're not sure whether something should be a notice, write it first in regular text
+and then decide if a notice is needed.
+
+Don't use too many notices. When you use multiple notices on a page, they begin to lose their
+visual distinctiveness. See if you can convey the information in a different way. This is
+especially true if you have two (or more) notices in a row.
+
+Where possible, avoid grouping two or more notices together. If you find it
+necessary to do so (for example, a *note* with a *caution* inside it, or several
+*warnings* one after another), consider reorganizing the content.
+
+<a id="notices--pick-a-notice-type"></a>
+
+#### Pick a notice type
+
+The following is a list of commonly used notices.
+
+<a id="notices--note"></a>
+
+**Note**
+    An ordinary aside or tip. Provides information that is useful but not critical to the reader.
+    For example, "Generating excessive amounts of traffic to external systems can resemble a
+    denial-of-service attack." For more information, see
+    [when to use](#notices--when-to-use-a-note-notice-type)
+    and [when not to use](#notices--when-not-to-use-a-note-notice-type) a note notice type.
+<a id="notices--caution"></a>
+
+**Caution**
+    Tells the reader to proceed carefully. For example, "We don't recommend using a
+    broad `0.0.0.0/0` range that would allow all traffic."
+<a id="notices--warning"></a>
+
+**Warning**
+    Stronger than a *caution* notice; it means "Don't do this" or that this step might be
+    irreversible, such as leading to permanent data loss. If a reader doesn't heed the warning, they
+    can lose money, lose work, or open themselves to a security breach. For example, "Don't put a
+    password on the command line; doing so is a security risk."
+<a id="notices--success"></a>
+
+**Success**
+    Describes a successful action or an error-free status. Used only in interactive or dynamic
+    content; don't use this notice type in ordinary static pages. For example, "You've successfully
+    deployed an application to GKE."
+
+<a id="notices--when-to-use-a-note-notice-type"></a>
+
+#### When to use a *note* notice type
+
+Create a *note* when all of the following are true:
+
+- The information you're sharing is *relevant* but not
+  *necessary* to what the reader is doing right now. If the reader skips
+  the information, they'll still succeed.
+- Interrupting the reader at this point is not an obstacle to the reader. For example, your
+  *note* isn't suggesting an alternative that leads the reader down a
+  different path.
+- The information is not part of the flow of what you're writing—it's not just
+  a continuation, a result, or a pointer to additional information.
+
+<a id="notices--when-not-to-use-a-note-notice-type"></a>
+
+#### When not to use a *note* notice type
+
+- Don't use *notes* for [cross-references](#cross-references).
+- Don't use *notes* to tell the reader about prerequisites or about
+  steps they should have taken earlier. Information like this should precede the
+  step.
+- Don't make a full procedural step into a *note*.
+- Don't use *notes* to provide information that's necessary for the
+  reader to succeed.
+- Don't use *notes* for information that's in flow with the preceding
+  text. For example, don't use a *note* to state expected results or to
+  include information that simply describes what precedes.
+
+<a id="notices--examples"></a>
+
+#### Examples
+
+Use whatever visual presentation for notices is standard for your site.
+
+If you're writing in HTML and your site doesn't specify what HTML to use for
+notices, we recommend using HTML code similar to the following example:
+
+```
+<aside class="note"><b>Note:</b> All VPC networks include firewall
+rules.</aside>
+```
+
+> [!NOTE]
+> **Note**: All VPC networks include firewall
+> rules.
+
+
+
+> [!CAUTION]
+> **Caution**: We don't recommend using a subnet that's part of a dynamic
+> route.
+
+
+
+> [!WARNING]
+> **Warning**: Do not manually edit or delete generated
+> table entries.
+
+
+
+> [!TIP]
+> **Success**: You've successfully created a Compute Engine instance.
+
+---
+
+<a id="numbers"></a>
+
+### Numbers
+
+*Source: <https://developers.google.com/style/numbers>*
+
+> [!NOTE]
+> For information about formatting quantities like 10 MB, see
+> [Units of measurement](#units-of-measure).
+
+<a id="numbers--ordinal-numbers"></a>
+
+#### Ordinal numbers
+
+Spell out all ordinal numbers in text.
+
+Recommended: first, fifth, twelfth,
+forty-third
+
+Not recommended: 1st, 5th, 12th, 43rd
+
+<a id="numbers--numbers-as-words"></a>
+
+#### Numbers as words
+
+This section covers when to spell out numbers as words.
+
+If it's important to have the number and associated noun together on the same line, use
+a nonbreaking space between the number and the noun.
+
+In general, spell out the following:
+
+- Numbers from zero through nine, except as noted in
+  [Numbers as numerals](#numbers--numbers-as-numerals).
+
+  Recommended:
+  two-day total
+
+  Recommended: four options
+
+  Recommended: five minutes
+
+  Recommended: nine developers
+- A number that starts a sentence.
+
+  Recommended: Fifteen
+  directories are created.
+
+  In some cases it's better to rearrange the sentence so that the number
+  appears later.
+
+  Recommended: In
+  general, avoid sending files larger than 164 MB as attachments.
+
+  Not recommended: 164 MB
+  is generally considered too large a file to send as an attachment.
+
+  **Exception**: It's okay, but non-optimal, to begin a
+  sentence with a four-digit year.
+- A number that is followed by a numeral.
+
+  Recommended: This
+  procedure creates fifteen 100,000-byte files.
+
+  *But*
+
+  Recommended: This
+  procedure creates 15 of the 100,000-byte files.
+- Indefinite and casual numbers.
+
+  Using words like *millions* or *billions* is fine for approximate numbers. For
+  precise numbers, use numerals.
+
+  Recommended: You
+  can specify thousands of combinations.
+
+  Recommended: The
+  API might return a list of a million songs.
+
+<a id="numbers--numbers-as-numerals"></a>
+
+#### Numbers as numerals
+
+This section covers when to use numerals to write numbers.
+
+If it's important to have the number and associated noun together on the same line, use
+a nonbreaking space between the number and the noun.
+
+In general, use numerals for the following:
+
+- Numbers 10 and greater.
+
+  Recommended: The link expires in 24
+  hours.
+
+  Recommended: 18 years old
+
+  Recommended: 27 minutes
+
+  Recommended: 728 shipments
+
+  Recommended: 18,000,000 users
+
+  Recommended: 10 chapters
+
+  Recommended: 102 degrees
+
+  **Exceptions**: Always use numerals for the following items, even if
+  they're less than 10:
+
+  - Version numbers.
+
+    Recommended:
+    version 3
+  - Technical quantities, such as amounts of memory, amounts of disk
+    space, numbers of queries, or usage limits.
+
+    Recommended: 6 queries per second
+
+    Recommended: 50 Mbps
+
+    Recommended: 128 bits
+  - Page numbers.
+  - Chapter numbers, sections, pages, and so on.
+  - Step numbers. Avoid referring to step numbers whenever possible,
+    but in edge cases where you have no choice or it makes the most sense,
+    use the numeral.
+  - Prices.
+  - Numbers without units, such as numbers used in mathematical
+    expressions.
+  - Numbers less than 10 when they appear in the same sentence with
+    numbers greater than 9.
+
+    Recommended: The
+    menu contains 15 options but 6 of them are deselected.
+- Negative numbers.
+- Most [fractions](#numbers--fractions).
+- [Percentages](#numbers--percentages).
+- [Dimensions](#numbers--dimensions).
+- Numbers containing decimal points.
+  - Treat decimal numbers as plural even when less than or equal to 1.0.
+
+    Recommended: 1.0 inches
+  - For decimal numbers less than one, place a zero in front of the decimal point.
+
+    Recommended: 0.3 inches
+- Measurements.
+
+  Recommended: 8 pixels
+- [Numbers in a range](#numbers--ranges-of-numbers).
+
+<a id="numbers--numbers-as-roman-numerals"></a>
+
+#### Numbers as Roman numerals
+
+In general, avoid using Roman numerals when possible. Instead, use Arabic numerals because they
+are easier to scan.
+
+You can use Roman numerals for
+[sub-steps in numbered procedures](#procedures--sub-steps-in-numbered-procedures).
+
+<a id="numbers--fractions"></a>
+
+#### Fractions
+
+Express fractions as decimal numbers, when possible.
+
+If you must express fractions as words, connect the numerator and
+denominator with a hyphen unless one of them is already hyphenated.
+
+Recommended: 0.75
+
+Recommended: one and one-half
+
+Recommended: two-fifths
+
+Recommended: five sixty-fourths
+
+<a id="numbers--percentages"></a>
+
+#### Percentages
+
+In general, use numerals and the percent sign (%), without a space between them.
+
+Recommended: 40%
+
+**Exception**: If the percentage starts the sentence, then spell out both
+the number and the word *percent*.
+
+Recommended: Forty
+percent of the files
+
+<a id="numbers--ranges-of-numbers"></a>
+
+#### Ranges of numbers
+
+Use a hyphen with no space on either side of it. Do not use an
+en dash (`&ndash;`).
+
+Recommended:
+2012-2016
+
+For more information, see the following:
+
+- [Ranges of numbers with units](#units-of-measure--ranges)
+- [Range of numbers](#hyphens--number-range)
+
+<a id="numbers--suspended-hyphens"></a>
+
+#### Suspended hyphens
+
+When two or more hyphenated compounds that start with numbers modify the same
+word, use [suspended hyphens](#hyphens--suspended-hyphens).
+
+Recommended: You can set up the system to
+scan for new files at one-, two-, or three-hour intervals.
+
+<a id="numbers--currency"></a>
+
+#### Currency
+
+Make sure that it's clear what country's currency you are describing. For more information, see
+the [currency](#units-of-measure--currency) section in Units of measurement.
+
+For US dollars, use a comma to delineate the thousands place of whole
+currency. Use a period to delineate whole currency and fractions of currency.
+Always include the dollar sign ($) at the beginning of the currency. Do
+not use any punctuation or spaces to the right of the decimal.
+
+Recommended: The price is $0.006653 per
+vCPU hour.
+
+Not recommended: The price is $0.006,653
+per vCPU hour.
+
+Recommended: $10,000 in fees is out of
+reach for many developers.
+
+Not recommended: $10 000 in fees is out
+of reach for many developers.
+
+<a id="numbers--punctuating-numbers"></a>
+
+<a id="numbers--commas-and-decimal-points-in-numbers"></a>
+
+#### Commas and decimal points in numbers
+
+Use commas and decimal points in accordance with standard American number-formatting.
+
+Specifically: in numbers four or more digits long, use commas to set off
+groups of three digits, counting leftward from the decimal point, in the
+standard American style. For long decimal numbers, do not use any digit-group separators to the
+right of the decimal point.
+
+> [!NOTE]
+> **Note**: Even though the
+> [International
+> System of Units](https://www.nist.gov/pml/weights-and-measures/metric-si/si-units) (SI) uses a thin space as a digit group separator, we use a comma, which is
+> the most common digit group separator used in the US.
+
+Use a period for a decimal point, also in the standard American style.
+
+| Recommended | Not recommended |
+| --- | --- |
+| The limit is 1,532,784 bytes per day. | The limit is 1532784 bytes per day. |
+| The API supports up to 2,000 vertices. | The API supports up to 2000 vertices. |
+| $0.031611/vCPU hour | $0.031 611/vCPU hour |
+
+> [!NOTE]
+> **Note**: Even though in some scientific writing,
+> four-digit numbers don't use commas, our style is to use a comma for a
+> four-digit number.
+
+For more information about decimal points and digit group separators, see Wikipedia's [decimal mark](http://wikipedia.org/wiki/Decimal_mark) entry.
+
+<a id="numbers--dimensions"></a>
+
+#### Dimensions
+
+Use numerals for dimensions.
+
+Use a lowercase *x* between the numerals in the dimensions, with no space between
+the numerals and the *x*.
+
+Recommended: 192x192
+
+Not recommended: 192 x 192
+
+<a id="numbers--exponents"></a>
+
+#### Exponents
+
+Use
+[standard mathematical notation](https://wikipedia.org/wiki/Exponentiation).
+Don't put a space between the base and the exponent.
+
+Recommended: 23
+
+<a id="numbers--real-world-implication"></a>
+
+<a id="numbers--accompany-numerical-concepts-with-real-world-practical-implications"></a>
+
+#### Accompany numerical concepts with real-world practical implications
+
+Accompany numerical concepts with real-world practical implications to provide tangible meaning.
+For example, if using a feature incurs additional fees, add a link to pricing calculator.
+
+<a id="numbers--math-concepts"></a>
+
+<a id="numbers--mathematical-notation-and-visuals"></a>
+
+#### Mathematical notation and visuals
+
+For general guidance on formatting mathematical notation, such as equations and variables, see
+[Mathematical notation](#mathematical-notation).
+
+---
+
+<a id="paragraph-structure"></a>
+
+### Paragraphs
+
+*Source: <https://developers.google.com/style/paragraph-structure>*
+
+Break up your paragraphs to aid in the scannability of the page and to avoid
+walls of text. Readers scan for information and read on different devices with
+different screen sizes. Each paragraph should address a single idea in the
+fewest words and in the fewest sentences possible.
+
+Don't make sentences longer in order to limit the number of sentences in a
+paragraph. Use shorter sentences and paragraphs.
+
+A paragraph longer than 5 or 6 sentences is often an indication that the
+paragraph is trying to convey too much information. If so, break the paragraph
+into smaller paragraphs or remove some content. However, don't break paragraphs
+up if they contain a single idea. It's OK to have a paragraph with one sentence,
+and it can be OK if it's longer than 6 sentences as long as it's still about one
+idea.
+
+<a id="paragraph-structure--put-critical-information-first"></a>
+
+#### Put critical information first
+
+Similarly to putting the most important information first in a sentence, put
+the most important information first in a paragraph. Don't hide the key point of
+a paragraph at the end of the paragraph. Readers don't read every word.
+
+<a id="paragraph-structure--format-paragraphs"></a>
+
+#### Format paragraphs
+
+Left-align text for readability. Don't center, full-justify, or right-align
+text.
+
+Don't force line breaks (hard returns) within sentences and paragraphs. Line
+breaks might not work well in resized windows, across different devices, or with
+enlarged text.
+
+---
+
+<a id="phone-numbers"></a>
+
+### Phone numbers
+
+*Source: <https://developers.google.com/style/phone-numbers>*
+
+This page describes how to use and format phone numbers in technical documentation. This page
+doesn't provide guidance on how to enter or use phone numbers in Google or third-party products.
+If you need information about entering phone numbers in a specific product, consult the
+product documentation or contact product support.
+
+<a id="phone-numbers--phone-numbers-in-examples"></a>
+
+#### Use example phone numbers
+
+Most phone numbers in our documentation are examples. To show an example phone number, use a US
+number in the range 800‑555‑0100 through 800‑555‑0199. That range is
+reserved for use in examples and in fiction.
+
+Never use a real phone number in examples.
+
+<a id="phone-numbers--format-phone-numbers"></a>
+
+#### Format phone numbers in HTML or Markdown
+
+To ensure that a phone number is displayed on the same line, use a nonbreaking hyphen
+(`&#8209;`) where appropriate in HTML or Markdown.
+
+<a id="phone-numbers--example"></a>
+
+##### Example
+
+415‑555‑0132
+
+<a id="phone-numbers--html"></a>
+
+##### HTML
+
+`415&#8209;555&#8209;0132`
+
+<a id="phone-numbers--markdown"></a>
+
+##### Markdown
+
+`415&#8209;555&#8209;0132`
+
+<a id="phone-numbers--phone-numbers-america"></a>
+
+#### Format North American phone numbers
+
+To format a real phone number in the US, Canada, and other [NANP](https://wikipedia.org/wiki/North_American_Numbering_Plan)
+(North American Numbering Plan) countries, use a nonbreaking hyphen to separate the area code,
+three-digit exchange code, and four-digit number.
+
+Recommended: 415‑555‑0132
+
+<a id="phone-numbers--international-phone-numbers"></a>
+
+#### Format international phone numbers
+
+To format a real phone number in non-NANP countries, include the country and area
+codes. Insert a plus sign
+immediately before the country code (no space); the plus sign stands in for a
+prefix known as an *exit code*, which lets you dial out of a country. Each
+country has a different exit code.
+
+For more information, see the [ITU document about
+standardized formatting for phone numbers](https://www.itu.int/rec/T-REC-E.123-200102-I/en).
+
+Recommended: +1‑415‑555‑0132
+
+<a id="phone-numbers--extensions"></a>
+
+#### Format phone numbers that include an extension
+
+To specify a phone extension, follow the phone number with the word *extension*, and then
+specify the extension number.
+
+Recommended: 415‑555‑0132, extension 987
+
+---
+
+<a id="procedures"></a>
+
+### Procedures
+
+*Source: <https://developers.google.com/style/procedures>*
+
+A procedure is a sequence of numbered steps for accomplishing a task. For information about
+lists of items that aren't part of a procedure, see the
+[Lists](#lists) page.
+
+<a id="procedures--introductory-sentences"></a>
+
+#### Introductory sentences
+
+In most cases, introduce a procedure with an introductory sentence. This
+introductory sentence should provide context to the reader that isn't part of
+the section heading. Don't simply repeat the heading: if the heading explains
+what the procedure is, and no additional context is needed, then don't
+include an introductory statement.
+
+The sentence can end with a colon or a period. Use a colon if it immediately
+precedes the procedure. Use a period if there's more material (such as a
+note paragraph) between the introduction and the procedure.
+
+You can introduce a procedure with an imperative statement. Don't introduce a procedure with
+a partial sentence that's completed by the numbered steps.
+
+Recommended: To customize the buttons,
+follow these steps:
+
+Also recommended: Customize the buttons:
+
+Also recommended: To customize the buttons, do the following:
+
+Not recommended: To customize the
+buttons:
+
+For more information about introducing lists, see [Lists](#lists--introductory-sentences-for-lists).
+
+<a id="procedures--single-step-procedures"></a>
+
+#### Single-step procedures
+
+When a procedure consists of only one step, write the step in one sentence and format it as a
+[bulleted list](#lists--numbered-lettered-bulleted-lists).
+
+Recommended:
+
+- To clear (flush) the entire log, click **Clear logcat**.
+
+Not recommended:
+
+To clear (flush) the entire log, follow this step:
+
+1. Click **Clear logcat**.
+
+Also not recommended:
+
+To clear (flush) the entire log, follow this step:
+
+- Click **Clear logcat**.
+
+<a id="procedures--sublists"></a>
+
+<a id="procedures--sub-steps-in-numbered-procedures"></a>
+
+#### Sub-steps in numbered procedures
+
+In a numbered procedure, sub-steps are labeled with lowercase letters, and
+sub-sub-steps get lowercase Roman numerals.
+
+When a step has sub-steps, treat the step like an [introductory sentence](#procedures--introductory-sentences): put a colon or a
+period at the end of the step, as appropriate.
+
+For more information about lists, see [Lists](#lists--introductory-sentences-for-lists).
+
+Recommended:
+
+1. To add a VM instance, do the following:
+   1. Click **Create instance**.
+   2. For **Name**, enter a name for the VM instance, and then do the following:
+      1. For **Region**, specify where you want to deploy the VM instance.
+      2. For **Machine type**, select an option.
+   3. Click **Create**.
+2. To connect to the VM instance by using SSH, click **SSH**.
+
+<a id="procedures--multiple-components"></a>
+
+<a id="procedures--order-of-multiple-components-in-a-step"></a>
+
+#### Order of multiple components in a step
+
+To document a complex procedural step, use the following order:
+
+1. Describe the action to take.
+2. List a command, if necessary.
+3. Explain any placeholders that are used in the command.
+
+   For more information, see
+   [Formatting placeholders](#placeholders).
+4. Explain the command in more detail, if necessary.
+5. List the output of the command, if necessary.
+
+   For more information, see
+   [Output from commands](#code-syntax--output).
+6. In a separate paragraph, explain
+   [the result of an
+   action](#procedures--steps-with-results-or-justifications), or any output, if necessary.
+
+The following example demonstrates the preceding order:
+
+1. Plan the Terraform deployment:
+
+   ```
+   terraform plan -out=NAME
+   ```
+
+   Replace `NAME` with the name of your Terraform plan.
+
+   The `terraform plan` command does the following:
+
+   1. Parses the Terraform configuration, building a list of resources to provision.
+   2. Refreshes the current state of resources already provisioned in Google Cloud.
+   3. Creates a plan to make the currently provisioned resources match the parsed
+      configuration.
+
+   The output is similar to the following:
+
+   ```
+     Plan: 26 to add, 0 to change, 0 to destroy.
+     ------------------------------------------------------------
+     This plan was saved to: NAME
+   ```
+
+   The output shows what resources to add, change, or destroy.
+
+<a id="procedures--multi"></a>
+
+<a id="procedures--multi-action-procedures"></a>
+
+#### Multi-action procedures
+
+In general, use one step for each action. However, you can combine small actions
+into one step [by using
+angle brackets](#ui-elements--term-menus) (`>`) for sequential menu selections.
+
+Recommended:
+
+1. Click **Next > Finish**.
+
+Also recommended:
+
+1. Click
+   **File > New > Document**.
+
+Don't make the steps too long. If they feel too long, consider splitting them
+into multiple steps.
+
+<a id="procedures--same-task"></a>
+
+<a id="procedures--multiple-procedures-for-the-same-task"></a>
+
+#### Multiple procedures for the same task
+
+In general, if there's more than one way to complete a task, then document
+one procedure that's accessible for all readers. If all methods are accessible, pick the shortest
+and simplest approach if possible. If you need to document multiple ways to complete a
+task, then separate them in different pages, headings, or tabs.
+
+The following guidelines can help you choose which procedure to document:
+
+- Choose a procedure that lets readers do all the steps by using only a keyboard.
+- Choose the shortest procedure.
+- Choose a procedure that uses a programming language that most of your
+  audience is familiar with.
+
+<a id="procedures--repetitive"></a>
+
+<a id="procedures--repetitive-procedures"></a>
+
+#### Repetitive procedures
+
+Avoid repeating procedures. Instead, reference those procedures and link to
+them.
+
+Recommended:
+
+1. Create a user as you did in the previous step.
+
+Also recommended:
+
+1. [Create a user as you did in the previous step.](#)
+
+<a id="procedures--optional-steps"></a>
+
+#### Optional steps
+
+For an optional step, at the beginning of the step, type *Optional*
+followed by a colon.
+
+Recommended:
+
+1. Optional: Type an arbitrary string ...
+
+Not recommended:
+
+1. (Optional) Type an arbitrary string ...
+
+For information about optional sections, see [Heading and title text](#headings--heading-and-title-text).
+
+<a id="procedures--specify-context"></a>
+
+<a id="procedures--steps-that-say-where-to-complete-a-task"></a>
+
+#### Steps that say where to complete a task
+
+Tell the reader where to complete an action—for example, in a
+particular tool or UI field—before you state the action.
+
+Recommended:
+
+1. In Google Docs, click
+   **File > New > Document**.
+2. In the Google Cloud console, go to the **Monitoring** page.
+
+Not recommended:
+
+1. Click **File >
+   New > Document** in Google
+   Docs.
+2. Go to the **Monitoring** page in the Google Cloud console.
+
+If a set of procedures is split across multiple headings, then in each
+procedure, restate where the reader completes the action. For example, if two
+procedures in a document take place in the console, then start both
+procedures with "In the console ..."
+
+<a id="procedures--steps-with-goals"></a>
+
+#### Steps with goals
+
+For some steps, it's useful to state the goal
+that the step accomplishes.
+
+When a step includes a goal, state the goal before the action. This
+structure helps readers understand and complete the step more easily.
+
+Recommended:
+
+1. To start a new document, click
+   **File >
+   New > Document**.
+
+Not recommended:
+
+1. Click **File >
+   New > Document** to start a
+   new document.
+
+Sometimes, the preceding format can imply that the required step is
+optional. In such cases, use the following format:
+
+Recommended:
+
+1. Start a new document: click
+   **File >
+   New > Document**.
+
+It's usually clear within the context of a procedure whether a step is
+required. In such cases, the "To ..." format is more natural than the colon
+format.
+
+To determine whether you need to use the colon format, consider how the
+goal of the step relates to the goal of the procedure. For example, in a
+procedure for creating a bar chart, a step with the goal "To create the
+chart" is clearly required. A step with the goal "To enhance the chart" is
+also unlikely to create confusion. But a step with the goal "To sort the
+data by date" might or might not be necessary. To clarify that the step
+isn't optional, use "Sort the data by date:" instead.
+
+<a id="procedures--steps-with-results"></a>
+
+<a id="procedures--steps-with-results-or-justifications"></a>
+
+#### Steps with results or justifications
+
+Some steps consist of an action along with a resulting reaction that helps
+the reader navigate to the next step. State the action first and the result
+second. Keep the result in the same paragraph as the action. But also consider whether you
+can avoid repetitiveness and overwhelming the reader with too much bolding of UI elements.
+
+Recommended:
+
+1. Click **Run**. The query results appear after the query runs.
+
+Recommended:
+
+1. Click **Enter**.
+2. In the **New file** dialog that appears, click **Next**.
+
+Not recommended:
+
+1. Click **Enter**. The **New file** dialog appears.
+2. In the **New file** dialog, click **Next**.
+
+For information about describing output, see
+[Output from commands](#code-syntax--output).
+
+Other steps benefit from including a justification for why the step is
+important. State the action first and the justification second.
+
+Recommended:
+
+1. Store the private key in a secure location. You need it later.
+
+<a id="procedures--more-guidelines-for-writing-procedures"></a>
+
+<a id="procedures--summary-of-guidelines-for-writing-procedures"></a>
+
+#### Summary of guidelines for writing procedures
+
+| Guidance | Recommended | Not recommended |
+| --- | --- | --- |
+| Make sure that the first sentence in a procedural step includes an imperative verb. | Clone the repository that contains the sample data. | You need the project ID later in this document. Retrieve the project ID. |
+| Use complete sentences. |  |  |
+| Use parallel structure and consistent verb form. | Download the service account key to your local machine. Click **More**, and then click **Download**. | Download the service account key to your local machine by clicking **More** and then clicking **Download** file. |
+| For an optional step, type *Optional:* as the first word of the step. | Optional: Type an arbitrary string... | (Optional) Type an arbitrary string... |
+| Set the context (such as a tool or an environment) in which the reader performs a procedure.  If there are multiple headings associated with a set of procedures, restate the context of the procedure in the first step, even if the context is the same as in the previous procedure. | In Cloud Shell, connect to the development cluster.  In the Google Cloud console, go to the **BigQuery** page. |  |
+| Write in the order that the reader needs to follow. State the location of the action before stating the action. | In Google Docs, click **File > New > Document**.  In the Google Cloud console, go to the **Monitoring** page. | Click **File > New > Document** in Google Docs.  Go to the **Monitoring** page in the Google Cloud console. |
+| State the purpose or goal of the action before stating the action. | To start a new document, click **File > New > Document**. | Click **File > New > Document** to start a new document. |
+| Don't use directional language to orient the reader, such as *above*, *below*, or *right-hand side*. This type of language doesn't work well for accessibility or for localization. If a UI element is hard to find, provide a screenshot.  For information about documenting icons, see [Buttons and icons](#ui-elements--buttons). | Click **Menu**.  In the preceding diagram,...  In the following diagram,... | Click the button with three lines.  In the above diagram, ...  In the diagram below, ... |
+| Don't use *please*. | To open a document, click **File > Open**. | To open a document, please click **File > Open**. |
+| Avoid using *run the following command* to introduce code. Instead, focus on what the command does. | In Cloud Shell, deploy the load generator:...  Define a firewall rule to allow internal traffic:... | In Cloud Shell, deploy the load generator by running the following command:...  Run the following command:... |
+| If the reader must press **Enter** after a step, then include that instruction as part of the step. | Click the search box, type `custom function`, and then press **Enter**. | 1. Click the search box and type `custom function`. 2. Press **Enter**. |
+| Don't include keyboard shortcuts. | Copy the command, and then paste it... | Press Ctrl+C, and then press Ctrl+V... |
+| When there's more than one way to do something, give only the best way. Giving alternate ways can confuse readers. |  |  |
+| If your procedure includes code samples, see how to format  [code samples](#code-samples). |  |  |
+| If your procedure includes commands, see how to format  [commands](#code-syntax--formatting-a-command). |  |  |
+| Ensure that the reader has the information that they need in order to prepare for the task ahead of time. Having information in advance supports task management, executive functioning, memory, and emotional regulation. | The following hardware and software are required:... |  |
+| Include as few steps as possible to complete the task. Limit interruptions in the path. |  |  |
+| Focus on one reader decision at a time. Separate each instruction by making each instruction a separate list item. |  |  |
+
+---
+
+<a id="tables"></a>
+
+### Tables
+
+*Source: <https://developers.google.com/style/tables>*
+
+In many contexts, tables are the best way to represent sets of related pieces of data. However,
+in some contexts, other approaches are better choices.
+
+<a id="tables--list-or-table"></a>
+
+#### List or table?
+
+Tables and lists are both ways to present a set of similarly structured
+items; sometimes it's not obvious when to choose one presentation over the
+other. To decide which presentation to use, consult the following table:
+
+| Item type | Example | How to present |
+| --- | --- | --- |
+| Each item is a single unit. | A list of programming language names, or a list of steps to follow. | Use a [numbered list, lettered list, or bulleted list](#lists--numbered-lettered-bulleted-lists). |
+| Each item is a pair of pieces of related data. | A list of term/definition pairs. | Use a [description list](#lists--description-lists) (or, in some contexts, a table). |
+| Each item is three or more pieces of related data. | A set of parameters, where each parameter has a name, a data type, and a description. | Use a table. |
+
+<a id="tables--places-not-to-use-tables"></a>
+
+##### Places not to use tables
+
+- Don't use tables to lay out a page; use your site's standard CSS instead.
+- Usually if you have only one row of material, a table isn't the best
+  choice for how to present it. But in some contexts (especially for consistency
+  of layout in reference documentation), it might be.
+- If you have only one column in your table, turn the table into a list.
+- Don't use tables to lay out code snippets.
+- Don't use tables to lay out long one-dimensional lists in multiple
+  columns. For example, if you have a long list of function names, don't try to
+  save space by splitting the list in half and presenting the two halves as a
+  two-column table. Use tables only to present two-dimensional data—that is,
+  material that semantically makes sense to display in rows and columns.
+- Avoid tables in the middle of a numbered procedure.
+
+<a id="tables--multiple-paragraphs"></a>
+
+<a id="tables--multi-paragraph-table-cells"></a>
+
+#### Multi-paragraph table cells
+
+Any table cell can contain more than one paragraph.
+
+To create multiple paragraphs, use the `p` element rather
+than using the `br` element. (The HTML specification
+describes which uses of the [`br`
+element](https://html.spec.whatwg.org/multipage/semantics.html#the-br-element) are legitimate and which aren't.)
+
+Example of a table with some cells that contain more than one paragraph:
+
+| Attribute name | Type | Description |
+| --- | --- | --- |
+| `href` | HTML | Defines the URL for a link.  For example, go to the `<a href="https://www.google.com">Google Search</a>` page. |
+| `src` | HTML | Defines the path of the image to be displayed.  For example, `<img src="kitten.jpg">`. |
+
+<a id="tables--table-introductions"></a>
+
+<a id="tables--introductory-sentences-for-tables"></a>
+
+#### Introductory sentences for tables
+
+Introduce tables with a complete sentence that describes the purpose of the table because not all
+screen readers preannounce tables. The introductory sentence can end with a colon or a period;
+usually a colon if it immediately precedes the table, and usually a period if there's more material
+(such as a note paragraph) between the introduction and the table.
+
+Recommended: Change the environment variables
+to values for your deployment, as listed in the following table:
+
+For more information, see the
+[Tables](#accessibility--tables) section of the "Accessibility" page.
+
+<a id="tables--table-placement"></a>
+
+#### Table placement
+
+- When introducing a table, use a complete sentence and try to refer to the
+  table's position, using a phrase like *the following table* or *the preceding table*.
+- Don't put a table in the middle of a sentence.
+- Avoid using footnotes when possible. If your table does refer to footnotes, place them
+  immediately following the table. For more information, see
+  [Footnotes](#footnotes).
+
+<a id="tables--table-captions"></a>
+
+#### Table captions
+
+If your document contains only one table, the table doesn't need a caption.
+However, be sure to place the table adjacent to the text that refers to it.
+
+If your document contains more than one table in fairly close proximity to
+each other, include a caption for each one, using a [`caption`
+element](https://html.spec.whatwg.org/multipage/tables.html#the-caption-element) as the first child of the `table` element. Start the
+caption with a number, in the form "<b>Table NUMBER.</b>
+DESCRIPTION". Use sentence case for the caption, but don't place a
+period at the end.
+
+When referring to the table from text, refer to it by its number—for example,
+*... as shown in table 2*. Do not capitalize *table* unless it starts a sentence.
+
+Your site's CSS determines the styling and placement of the caption.
+
+Recommended:
+
+```
+<table>
+  <caption><b>Table 1.</b> Prehistoric birds</caption>
+  ...
+</table>
+```
+
+<a id="tables--table-formatting"></a>
+
+#### Table formatting
+
+- Don't add styling to the table element.
+- Don't apply a visual style such as a different font, font color, or background color to convey a
+  header row or column by itself. Use the `th` element to semantically mark up headers in
+  tables.
+- Don't merge cells. Don't use `colspan` or `rowspan` attributes.
+- Sort rows in a logical order, or alphabetically if there is no logical order.
+- If the table is long or complicated—for example, with multiple header rows or columns—consider
+  splitting it into multiple tables.
+- Don't present new information in tables through images or symbols alone; always provide a
+  descriptive `alt` attribute for the image or symbol. For more information, see
+  [Alt text](#images--alt-text).
+
+<a id="tables--table-column-heads"></a>
+
+#### Table column heads
+
+- Use sentence case.
+- Write concise headings.
+- Don't end with punctuation, including a period, an ellipsis, or a colon.
+- Use table headings for the first column and the first row only. Use the
+  [`th` element](https://www.w3.org/TR/2014/REC-html5-20141028/tabular-data.html#the-th-element).
+- Include the [`scope` attribute](https://www.w3.org/TR/WCAG20-TECHS/H63.html) as appropriate for accessibility.
+
+<a id="tables--responsive-tables"></a>
+
+#### Responsive tables
+
+Where possible, use table CSS that adapts to different viewport sizes.
+
+<a id="tables--table-linking"></a>
+
+<a id="tables--link-to-tables"></a>
+
+#### Link to tables
+
+Where possible, avoid linking to tables; instead, refer to them by table number.
+
+---
+
+<a id="units-of-measure"></a>
+
+### Units of measurement
+
+*Source: <https://developers.google.com/style/units-of-measure>*
+
+Put a nonbreaking space (`&nbsp;`) between the number and the unit.
+
+<a id="units-of-measure--spaces"></a>
+
+<a id="units-of-measure--spaces-in-units-of-measurement"></a>
+
+#### Spaces in units of measurement
+
+For most units of measurement, when you specify a number with the unit, use a
+nonbreaking space between the number and the unit. This guidance applies in both
+HTML and Markdown.
+
+For guidance about when to spell out units, see the
+[Abbreviations](#abbreviations--spelling-out) page.
+
+For guidance about whether to hyphenate, see the [Hyphens](#hyphens--compounds)
+page.
+
+Recommended: `64&nbsp;GB` (64 GB)
+
+Recommended: `25&nbsp;mm` (25 mm)
+
+Recommended: a 128-bit system
+
+Not recommended: `64 GB`
+
+Not recommended: 64GB
+
+For more information about making abbreviations plural, see [Plural abbreviations](#pluralization--making-abbreviations-plural).
+
+However, when the unit of measure is money or percent or degrees of an angle,
+don't use a space. For more information, see [Currency](#units-of-measure--currency).
+
+Recommended: $10
+
+Recommended: £25
+
+Recommended: 65%
+
+Recommended: 180°
+
+For degrees of temperature, include a nonbreaking space between the number and the degree symbol.
+Don't use a space between the degree symbol (`&deg;`) and the temperature scale
+(*F* or *C*).
+
+<a id="units-of-measure--example"></a>
+
+##### Example
+
+50 °C
+
+<a id="units-of-measure--html"></a>
+
+##### HTML
+
+`50&nbsp;&deg;C`
+
+<a id="units-of-measure--markdown"></a>
+
+##### Markdown
+
+`50&nbsp;&deg;C`
+
+For Kelvin temperatures, leave out the degree symbol but use a nonbreaking space before the
+*K*.
+
+<a id="units-of-measure--example_1"></a>
+
+##### Example
+
+300 K
+
+<a id="units-of-measure--html_1"></a>
+
+##### HTML
+
+`300&nbsp;K`
+
+<a id="units-of-measure--markdown_1"></a>
+
+##### Markdown
+
+`300&nbsp;K`
+
+When a number and unit of measurement combine to modify a noun, don't hyphenate unless
+the hyphen is needed for clarity.
+
+Recommended:
+`200&nbsp;GB disk` (200 GB disk)
+
+<a id="units-of-measure--ranges"></a>
+
+<a id="units-of-measure--ranges-of-numbers-with-units"></a>
+
+#### Ranges of numbers with units
+
+In a range of numbers, repeat the unit for each number. *Unit* includes both symbols (like
+the degree symbol (º)) and abbreviations (like *MB* for megabytes) but not nouns
+(like *file*). For more information, see
+[Range of numbers](#hyphens--number-range).
+
+Use the word *to* between the numbers, rather than a hyphen. A hyphen
+can be misinterpreted as a subtraction sign.
+
+Recommended: -40 °C to 85 °C
+
+Not recommended: -40-85 °C
+
+<a id="units-of-measure--multiplied-units"></a>
+
+<a id="units-of-measure--hyphens-with-multiplied-units"></a>
+
+#### Hyphens with multiplied units
+
+When the components of a unit of measurement are multiplied by each other,
+hyphenate them.
+
+Recommended: 5 vCPU-hours
+
+Recommended: 40 person-hours
+
+<a id="units-of-measure--using-k"></a>
+
+<a id="units-of-measure--use-k-to-indicate-thousands"></a>
+
+#### Use *k* to indicate thousands
+
+In some contexts, it might be appropriate to indicate thousands of something by
+following a number with a lowercase *k*. If you do that, then follow these
+guidelines:
+
+- Don't put a space between the number and *k*.
+- Add a noun to indicate what the number measures, and to make clear that
+  you're not using *k* as an abbreviation for *kilobytes*.
+
+Recommended: On this plan, you are
+limited to 55k download operations and 20k upload operations per day.
+
+<a id="units-of-measure--currency"></a>
+
+#### Currency
+
+If you're writing about monetary amounts, make sure that the reader knows what
+currency you're referring to. For example, the dollar sign—the *$*
+symbol—can refer to US dollars, Canadian dollars, Mexican pesos, and several
+other currencies.
+
+If there's any possibility of ambiguity, use a currency indicator before
+the amount. For details, see section 9.20 and following in the Chicago
+Manual of Style, 17th edition.
+
+Recommended: US$10
+
+<a id="units-of-measure--rates"></a>
+
+#### Rates
+
+Use *per* instead of the division slash (/) when space permits.
+It's OK to use the division slash when space is limited,
+such as in a table with small cells.
+
+Shorten *per* to *p* only for well-established abbreviations for
+rate units, such as *Gbps* for *gigabits per second* or
+*MBps* for *megabytes per second*.
+
+Recommended: requests per day
+
+Not recommended: requests/day
+
+Recommended: Gbps
+
+Not recommended: Gb/s
+
+<a id="units-of-measure--byte-units"></a>
+
+#### Decimal and binary units
+
+Use the same system to measure bytes as the technology that you're documenting.
+Don't use *MB* if you mean *MiB*, or *GB* if you mean *GiB*. The following
+table lists common types of
+[decimal and binary units](https://en.wikipedia.org/wiki/Byte#Multiple-byte_units):
+
+| Decimal units | Binary units |
+| --- | --- |
+| kB (kilobyte, or 1000 bytes) | KiB (kibibyte, or 1024 bytes) |
+| MB (megabyte, or 10002 bytes) | MiB (mebibyte, or 10242 bytes) |
+| GB (gigabyte, or 10003 bytes) | GiB (gibibyte, or 10243 bytes) |
+
+For more information about abbreviating measurement terms, see
+[When to spell out a term](#abbreviations--spelling-out).
+
+<a id="units-of-measure--more-resources"></a>
+
+#### More resources
+
+- [Mathematical notation](#mathematical-notation)
+- [Numbers](#numbers)
+
+---
+
+## Linking
+
+<a id="cross-references"></a>
+
+### Cross-references and linking
+
+*Source: <https://developers.google.com/style/cross-references>*
+
+In general, cross-references link to nonessential information that adds to
+the reader's understanding.
+
+When used well, cross-references help readers navigate and understand
+documentation. But cross-references can easily become disruptive. The guidelines
+on this page help you to minimize disruption while providing cross-references
+that help your readers.
+
+<a id="cross-references--choose-links"></a>
+
+#### Choose links selectively
+
+Be selective about which links you include on a page. Each link creates a
+decision for the reader, adding cognitive load. Each link is also a chance for
+the reader to leave the page and lose their place. When you include links,
+choose the most relevant destination.
+
+<a id="cross-references--context-on-page"></a>
+
+##### Provide context on the page
+
+When possible, provide help in context rather than linking elsewhere. For
+example, in the following situations, consider providing information on the
+page instead of linking:
+
+- Define a term.
+- Briefly explain a concept.
+- Provide a couple of steps.
+
+As a specific example, if you need readers to understand another product's
+software or standards, it's better to link to good documentation elsewhere
+than to try to thoroughly document another product's standards in our
+documentation. But if a few sentences of basic information is all your readers
+need, then it's better to provide that context and save your readers the trip
+outside of our documentation.
+
+<a id="cross-references--avoid-duplicate-links"></a>
+
+##### Avoid duplicate links
+
+Generally, within a given page, don't provide duplicate links to the same
+destination. Provide the link once in the location where it's most useful to
+the reader.
+
+It's OK to add a secondary link in situations such as the following:
+
+- You're linking to a particular section of another page.
+- Your page is very long and the duplicate links are far apart.
+- There are multiple entry points to the document that you're linking from.
+  For example, if a page contains a procedure section and a troubleshooting
+  section, then you might need to provide the same link in both of those
+  sections.
+
+<a id="cross-references--relevant-links"></a>
+
+##### Provide the most relevant link
+
+When you link, link to the most relevant page on a site. Link to the most
+relevant heading on a page. Avoid providing multiple links that do the same
+job.
+
+<a id="cross-references--third-party-sites"></a>
+
+##### Link to third-party sites
+
+Our documentation often relies on the reader knowing something about
+third-party standards or software. In such cases, it's better to provide a
+link rather than attempt to thoroughly document someone else's standards. But
+as with all links, when possible, provide brief information on the page
+instead of linking.
+
+<a id="cross-references--descriptive-link-text"></a>
+
+#### Write descriptive link text
+
+For the link text itself, use short, unique, descriptive phrases that provide
+context for the material that you're linking to.
+
+Effective link text helps to improve accessibility and scannability. Different
+readers experience links differently. For example, users of screen reader
+software often jump from one link to the next without reading the words in
+between. Other readers visually scan a document to find relevant links.
+
+Sometimes you have to rework a sentence to include a phrase that makes good
+link text.
+
+<a id="cross-references--link-text-options"></a>
+
+##### Two options for effective link text
+
+For your link text, use either the exact page title or a descriptive phrase, as
+described in the following sections.
+
+<a id="cross-references--page-titles"></a>
+
+###### Page titles as link text
+
+One option for effective link text is to match the link text to the page
+title or heading that you're referencing.
+
+For more information about how to capitalize the page title in a
+cross-reference, see
+[Capitalization in references to titles and headings](#capitalization--capitalization-in-references-to-titles-and-headings).
+
+Recommended: For more
+information, see
+[Load balancing and scaling](https://cloud.google.com/compute/docs/load-balancing-and-autoscaling).
+
+<a id="cross-references--descriptive-phrases"></a>
+
+###### Descriptive phrases as link text
+
+Another option for effective link text is to use a description of the
+destination page, capitalized as if it's part of the sentence.
+
+When you write a descriptive phrase as link text, help readers quickly
+determine whether the link is relevant to them:
+
+- Place important words at the beginning of the link text.
+- Don't use the same link text in the same document for different target
+  pages.
+- Keep link text short where possible.
+  Don't write lengthy link text such as a sentence or short paragraph.
+
+Recommended: You can use
+Cloud Scheduler and Cloud Functions to manage
+[task scheduling on Compute Engine](https://cloud.google.com/blog/products/gcp/reliable-task-scheduling-on-google-compute-engine).
+
+Not recommended: See
+[this blog post](https://www.blog.google/products/pixel/pixel-4/).
+
+<a id="cross-references--vague-link-text"></a>
+
+##### Avoid vague link text
+
+Write link text that makes sense without the surrounding text.
+Don't use phrases such as *this document*, *this article*, or *click here*.
+
+Recommended:
+For more information, see
+[Make headings into link targets](#headings-targets).
+
+Not recommended:
+Want more? [Click here!](#headings-targets)
+
+Not recommended:
+For more information,
+see [this document](#headings-targets).
+
+<a id="cross-references--urls"></a>
+
+##### Avoid URLs as link text
+
+In general, don't use a URL as link text. Instead, use the page title or a
+description of the page.
+
+Recommended:
+
+```
+For more information about protocols, see <a href="http://www.w3.org/Protocols/rfc2616/rfc2616.html">HTTP/1.1 RFC</a>.
+```
+
+Not recommended:
+
+```
+  See the HTTP/1.1 RFC at <a href="http://www.w3.org/Protocols/rfc2616/rfc2616.html">http://www.w3.org/Protocols/rfc2616/rfc2616.html</a>.
+```
+
+**Exception**: In some legal documents (such as some Terms of Service documents), it's
+okay to use URLs as link text.
+
+<a id="cross-references--abbreviations"></a>
+
+##### Include abbreviations in link text
+
+If the text includes an abbreviation in parentheses, include the long form
+and the abbreviation in the link text.
+
+Recommended: [Google Kubernetes Engine (GKE)](https://cloud.google.com/kubernetes-engine/docs)
+
+Not recommended: [Google Kubernetes Engine](https://cloud.google.com/kubernetes-engine/docs) (GKE)
+
+<a id="cross-references--commands"></a>
+
+##### Link to commands
+
+If the text includes a command or another element usually conveyed with
+code font, include the description of the code element with the link text,
+unless doing so is awkward or redundant. For more information about elements
+that appear in code font, see [Code in text](#code-in-text).
+
+Recommended: To create an
+instance with a custom hostname, run the `gcloud instances create`
+command with the
+[`--hostname` flag](https://cloud.google.com/compute/docs/instances/custom-hostname-vm#gcloud).
+
+Not recommended: To create
+an instance with a custom hostname, run the `gcloud instances create`
+command with the
+[`--hostname`](https://cloud.google.com/compute/docs/instances/custom-hostname-vm#gcloud)
+flag.
+
+Recommended: This service
+supports the `GET`, `HEAD`,
+and `OPTIONS` methods.
+
+Not recommended: This
+service supports the `GET` method,
+`HEAD` method, and
+`OPTIONS` method.
+
+<a id="cross-references--link-introductions"></a>
+
+#### Write link introductions ("For more information")
+
+When you dedicate a separate sentence to a cross-reference, introduce the
+cross-reference using consistent language—specifically, use the phrase "For more
+information, see..." or "For more information about..., see... ."
+
+Include the "about..." clause when the link text or surrounding context
+doesn't clearly indicate why you're referring the reader to this information.
+For more information, see the
+[Clarify the purpose of a link](#cross-references--clarify-purpose)
+section of this document.
+
+Don't use *on* instead of *about*.
+
+Use *see* to refer to links and cross-references. For more information, see
+[see](#word-list--see).
+
+Recommended: For more information, see
+[Load balancing and scaling](https://cloud.google.com/compute/docs/load-balancing-and-autoscaling).
+
+Recommended: For more information about
+task scheduling, see
+[Reliable task scheduling on Google Compute Engine](https://cloud.google.com/blog/products/gcp/reliable-task-scheduling-on-google-compute-engine).
+
+Not recommended: For more information on
+indexes, see [Manage indexes](https://cloud.google.com/firestore/docs/query-data/indexing).
+
+<a id="cross-references--clarify-purpose"></a>
+
+#### Clarify the purpose of a link
+
+Make sure that the surrounding context or the link text itself clearly
+indicates why you're referring the reader to this information. Make the
+explanation specific, but don't repeat the link text.
+
+If you're introducing a cross-reference with "For more information..."
+phrasing, then you can do this by adding an "about..." phrase. For more
+information, see the
+[Write link introductions](#cross-references--link-introductions) section
+of this document.
+
+Recommended: For more
+information about authentication and authorization, see
+[Using OAuth 2.0 to access Google APIs](https://developers.google.com/identity/protocols/OAuth2).
+
+Recommended: If your
+sample dump file is in a CSV, Avro, or Parquet file format, then
+[load the file to BigQuery and copy to Spanner](https://cloud.google.com/spanner/docs/load-sample-data) using reverse ETL.
+
+<a id="cross-references--explain-behavior"></a>
+
+#### Explain unexpected link behavior
+
+If a link goes to an unexpected destination or behaves in an unexpected way,
+then provide that context. The following are a few such situations:
+
+- **Links that download files and open emails.** If a link
+  downloads a file or opens an email, then make that clear in the link text, and
+  mention the file type.
+
+  Recommended: For more
+  information,
+  [download the security features PDF](https://www.example.com/security.pdf).
+
+  Recommended:
+
+  ```
+    <a href="mailto:support@example.com">send email to Technical Support</a>
+  ```
+- **Links to sections on the same page.** When you're
+  linking to another section on the same page, let the reader know that the link
+  takes you to a different section of the same page. Use a standard phrase to clue
+  readers in if you use an on-page link.
+
+  Recommended: For more
+  information, see the
+  [Write descriptive link text](#cross-references--descriptive-link-text)
+  section of this document.
+- **Links to sections on another page.** When you're linking
+  to a section heading on another page, use the same wording and formatting as you
+  do in a regular cross-reference.
+
+  If the title of the section that you're linking to is identical to a
+  title on the source page, add context to the cross-reference.
+
+  Recommended: For more information, see
+  [Create a table](https://cloud.google.com/bigtable/docs/managing-tables#create-table).
+
+  Recommended: For more information, see
+  [Install libraries](https://developers.google.com/style/cross-references#different-page)
+  in "Building new audiences based on existing customer lifetime value."
+- **Links that open in a new tab.** For more information, see the
+  [Open links in the current tab](#cross-references--current-tab) section of this
+  document.
+- **Links that go to a different domain or server.** For more
+  information, see the
+  [Don't use external link icons](https://developers.google.com/style/cross-references#external-link-icons) section
+  of this document.
+
+<a id="cross-references--current-tab"></a>
+
+#### Open links in the current tab
+
+Don't force links to open in a new tab or window. Let the reader decide how
+to open links.
+
+In the rare situation that a link needs to open in a new tab or window, let
+the reader know that the link opens differently than expected.
+
+Recommended:
+
+```
+<a href="/style/accessibility">Accessible content</a>
+```
+
+Recommended:
+
+```
+<a href="/style/accessibility" target="_blank">Accessible content (opens in a new tab)</a>
+```
+
+Not recommended:
+
+```
+<a href="/style/accessibility" target="_blank">Accessible content</a>
+```
+
+---
+
+<a id="headings-targets"></a>
+
+### Headings as link targets
+
+*Source: <https://developers.google.com/style/headings-targets>*
+
+This page discusses how to turn a heading into a link target by using an
+`id` attribute. For more information about how to format headings, see
+[Headings and titles](#headings).
+
+In some content management systems, anchors are automatically created for headings. However, you
+might want to add a *custom* anchor to a heading for several reasons:
+
+- You want to use an anchor that's shorter than the automatically generated anchor.
+- You want to use an anchor for content that might be frequently linked to. Adding a custom
+  anchor reduces the likelihood of breaking existing links if the heading text changes later.
+- You want to [revise a heading](#headings-targets--changing-an-anchor). If the
+  anchor for the heading is generated automatically, then the anchor changes when you revise
+  the heading, breaking existing links.
+
+<a id="headings-targets--adding-an-anchor"></a>
+
+<a id="headings-targets--add-a-custom-anchor"></a>
+
+#### Add a custom anchor
+
+<a id="headings-targets--html"></a>
+
+##### HTML
+
+To add an anchor to a heading in HTML, add a `section` element
+with an `id` attribute, or use an `a` element with
+a `name` attribute. For anchor text, use lowercase letters, and
+put hyphens between words. In the following, replace
+`ID_OF_ANCHOR` with your anchor text—for example,
+`introduction-to-everything`.
+
+```
+    <section id="ID_OF_ANCHOR"></section>
+  
+```
+
+Recommended:
+
+```
+  <section id="introduction-to-everything">
+  <h2>Introduction to everything</h2>
+  ...
+  </section>
+  
+```
+
+Recommended:
+
+```
+  <h2><a name="introduction-to-everything">Introduction to everything</a></h2>
+  
+```
+
+Recommended:
+
+```
+  <a name="introduction-to-everything"></a>
+  <h2>Introduction to everything</h2>
+  
+```
+
+Acceptable:
+
+```
+  <h2 id="introduction-to-everything">Introduction to everything</h2>
+  
+```
+
+<a id="headings-targets--markdown"></a>
+
+##### Markdown
+
+To add an anchor to a heading in Markdown, add the following code to the
+end of the line that the heading is on. For anchor text, use lowercase
+letters, and put hyphens between words. In the following, replace
+`ID_OF_ANCHOR` with your anchor text—for example,
+`conserve-habitat`.
+
+```
+{: #ID_OF_ANCHOR }
+```
+
+Recommended:
+
+```
+## Help conserve habitat for pollinators {: #help-conserve-habitat-for-pollinators }
+```
+
+Also recommended:
+
+```
+## Help conserve habitat for pollinators {: #conserve-habitat }
+```
+
+Acceptable:
+
+```
+## Help conserve habitat for pollinators {: id='conserve-habitat' }
+```
+
+Acceptable:
+
+```
+## Help conserve habitat for pollinators {: id="conserve-habitat" }
+```
+
+<a id="headings-targets--changing-an-anchor"></a>
+
+<a id="headings-targets--revise-a-heading"></a>
+
+#### Revise a heading
+
+If you revise a heading in a content management system where anchors are automatically created,
+you can create a custom anchor to avoid breaking existing links. If the heading already has a
+custom anchor, don't change the anchor unless it contains a term that you want to remove (such as
+a disrespectful term).
+
+To create the custom anchor, use the older ID string for the heading. You can find the ID
+string by inspecting the heading on the published page. For example, if you change a heading from
+*Introduction to some things* to *Introduction to everything*, then add a custom anchor
+that uses the older ID string and formatting.
+
+<a id="headings-targets--html_1"></a>
+
+##### HTML
+
+```
+<section id="introduction-to-some-things">
+<h2>Introduction to everything</h2>
+...
+</section>
+```
+
+<a id="headings-targets--markdown_1"></a>
+
+##### Markdown
+
+```
+## Introduction to everything {: #introduction-to-some-things }
+```
+
+If you need to change an existing custom anchor, you should check your content management
+system to update any links that use the old anchor. Inbound links that use the old anchor still
+reach the page but not the specific section or heading.
+
+---
+
+## Computer interfaces
+
+<a id="api-reference-comments"></a>
+
+### API reference code comments
+
+*Source: <https://developers.google.com/style/api-reference-comments>*
+
+When you're documenting an API, provide a complete API reference, typically
+generated from source code using document comments that describe all public
+classes, methods, constants, and other members.
+
+Use the basic guidelines in this document as appropriate for a given programming
+language. This document doesn't specify how to mark up document comments.
+
+For more information, see the following resources:
+
+- [AIP-192: Documentation](https://google.aip.dev/192)
+  in Google's API standards
+- [Inline API documentation](https://cloud.google.com/apis/design/documentation)
+  in the Google Cloud API design guide
+- The specific style guide for each programming language
+
+<a id="api-reference-comments--documentation_basics"></a>
+
+#### Documentation basics
+
+The API reference **must** provide a description for each of the following:
+
+- Every class, interface, struct, and any other similar member of the API (such
+  as union types in C++).
+- Every constant, field, enum, and typedef.
+- Every method, with a description for each parameter, the return value, and any
+  exceptions thrown.
+
+The following are **extremely strong suggestions**. In some cases, they don't
+make sense for a particular API or in a specific language, but in general,
+follow these guidelines:
+
+- On each unique page (for a class, interface, etc.), include a code sample
+  (~5-20 lines) at the top.
+- Put all API names, classes, methods, constants, and parameters in code font,
+  and link each name to the corresponding reference page. Most document
+  generators do this automatically for you.
+- Put string literals in code font, and enclose them in double quotation marks.
+  For example, XML attribute values might be `"wrap_content"` or `"true"`.
+- Make sure that the spelling of a class name in documentation matches the
+  spelling in code, with capital letters and no spaces (for example,
+  `ActionBar`).
+
+  - Don't make class names plural (`Intents`, `Activities`); instead, add a
+    plural noun (`Intent` objects, `Activity` instances). For more
+    information, see [Plural product and feature
+    names](#pluralization--plural-product-and-feature-names).
+  - However, if a class has a name that's a common term, you can refer to it
+    with the corresponding English word, in lowercase and *not* in code font
+    (activities, action bar).
+
+<a id="api-reference-comments--classes_interfaces_structs"></a>
+
+#### Classes, interfaces, structs
+
+In the first sentence of a class description, briefly state the intended purpose
+or function of the class or interface with information that can't be deduced
+from the class name and signature. In additional documentation, elaborate on how
+to use the API, including how to invoke or instantiate it, what some of the key
+features are, and any best practices or pitfalls.
+
+Many documentation tools automatically extract the first sentence of each class
+description for use in a list of all classes, so make the first sentence unique
+and descriptive, yet short. Additionally:
+
+- Don't repeat the class name in the first sentence.
+- Don't say "this class will/does ..."
+- Don't use a period before the actual end of the sentence, because some
+  document generators naively terminate the "short description" at the first
+  period. For example, some generators terminate the sentence if they see
+  *e.g.*, so use *for example* instead.
+
+The following example is the first sentence of the description for Android's
+[`ActionBar` class](http://developer.android.com/reference/android/app/ActionBar.html):
+
+> *A primary toolbar within the activity that may display the activity title,
+> application-level navigation affordances, and other interactive items.*
+
+<a id="api-reference-comments--members"></a>
+
+#### Members
+
+Make descriptions for members (constants and fields) as brief as possible. Be
+sure to link to relevant methods that use the constant or field.
+
+For example, here's the description for the `ActionBar` class's
+[`DISPLAY_SHOW_HOME`](http://developer.android.com/reference/android/app/ActionBar.html#DISPLAY_SHOW_HOME)
+constant:
+
+> *Show 'home' elements in this action bar, leaving more space for other
+> navigation elements. This includes logo and icon.*
+
+> *See also: `setDisplayOptions(int)`, `setDisplayOptions(int, int)`*
+
+<a id="api-reference-comments--methods"></a>
+
+#### Methods
+
+In the first sentence for a method description, briefly state what action the
+method performs. In subsequent sentences, explain why and how to use the method,
+state any prerequisites that must be met before calling it, give details about
+exceptions that may occur, and specify any related APIs.
+
+Document any dependencies (such as
+[Android permissions](http://developer.android.com/guide/topics/security/permissions.html))
+that are needed to call the method, and how the method behaves if such a
+dependency is missing (for example, "the method throws a
+[SecurityException](http://developer.android.com/reference/java/lang/SecurityException.html)"
+or "the method returns null").
+
+For example, here's the description for Android's
+[`Activity.isChangingConfigurations` method](http://developer.android.com/reference/android/app/Activity.html#isChangingConfigurations()):
+
+> *Checks whether this activity is in the process of being destroyed in order to
+> be recreated with a new configuration. This is often used in `onStop` to
+> determine whether the state needs to be cleaned up or if it's passed on to the
+> next instance of the activity using `onRetainNonConfigurationInstance`.*
+
+Use present tense for all descriptions—for example:
+
+- *Adds a new bird to the ornithology list.*
+- *Returns a bird.*
+
+<a id="api-reference-comments--description"></a>
+
+##### Description
+
+- If a method performs an operation and returns some data, start the description
+  with a verb describing the operation—for example:
+
+  - *Adds a new bird to the ornithology list and returns the ID of the new
+    entry.*
+- If it's a "getter" method and it returns a boolean, start with "Checks
+  whether ...."
+- If it's a "getter" method and it returns something other than a boolean,
+  start with "Gets the ...."
+- If it has no return value, start with a verb like one of the following:
+
+  - Turning on an ability or setting: "Sets the ...."
+  - Updating a property: "Updates the ...."
+  - Deleting something: "Deletes the ...."
+  - Registering a callback or other element for later reference:
+    "Registers ...."
+  - For a callback: "Called by ...." (Usually for a method that's named
+    starting with "on", such as `onBufferingUpdate`.) For example, "Called by
+    Android when ...." Then, later in the description: "Subclasses implement this
+    method to ...."
+- If it's a convenience method that constructs the class object, start with
+  "Creates a ...."
+
+<a id="api-reference-comments--parameters"></a>
+
+##### Parameters
+
+For parameter descriptions, follow these guidelines:
+
+- Capitalize the first word, and end the sentence or phrase with a period.
+- Begin descriptions of non-boolean parameters with "The" or "A" if possible:
+
+  - *The ID of the bird you want to get.*
+  - *A description of the bird.*
+- For boolean parameters that tell the API to do or not do something, state
+  what the API does if the parameter is true and if it's false. For example:
+
+  - *`enableCertificateValidation`: If true, validates the SSL certificate
+    before proceeding. If false, trusts the certificate without validating it.*
+- For boolean parameters that declare the already-established state of something
+  (rather than telling the API to do something), use the format "True if ...;
+  false otherwise." For example:
+
+  - *True if the zoom is set; false otherwise.*
+- In this context, don't put the words "true" and "false" in code font or
+  quotation marks.
+- For parameters with default behavior, explain what the behavior is for each
+  value or range of values, and then say what the default value is. Use the
+  format *Default:* to explain the default value.
+
+<a id="api-reference-comments--return_values"></a>
+
+##### Return values
+
+Be as brief as possible in the return value's description; put any detailed
+information in the class description.
+
+- If the return value is anything other than a boolean, start with "The ..."—for
+  example:
+
+  - *The bird specified by the given ID.*
+- If the return value is a boolean, use the format "True if ...; false
+  otherwise."—for example:
+
+  - *True if the bird is in the sanctuary; false otherwise.*
+
+<a id="api-reference-comments--exceptions"></a>
+
+##### Exceptions
+
+In languages where the reference generator automatically inserts the word
+"Throws", begin your description with "If ...":
+
+- *If no key is assigned.*
+
+Otherwise, begin with "Thrown when ...":
+
+- *Thrown when no key is assigned.*
+
+<a id="api-reference-comments--deprecations"></a>
+
+##### Deprecations
+
+When something is deprecated, tell the user what to use as a replacement. (If
+you track your API with version numbers, mention which version it was first
+deprecated in.)
+
+Only the first sentence of a description appears in the summary section and
+index, so put the most important information there. Subsequent sentences can
+explain why something is deprecated, along with any other information that's
+useful for a developer using your API.
+
+If a method is deprecated, tell the reader what to do to make their code work.
+
+<a id="api-reference-comments--examples"></a>
+
+###### Examples
+
+> *Deprecated. Use #CameraPose instead.*
+
+> *Deprecated. Access this field using the `getField` method.*
+
+---
+
+<a id="code-in-text"></a>
+
+### Code in text
+
+*Source: <https://developers.google.com/style/code-in-text>*
+
+In ordinary text sentences (as opposed to, say, [code samples](#code-samples)),
+use code font to mark up most things that have anything to do with code. Code
+font helps to clarify for your reader which text refers to an entity in these
+ways:
+
+- Signals to your reader that the text is meant to be entered
+  verbatim.
+- Shows where the boundaries of the text to enter are.
+- Clearly separates the entity from surrounding text.
+
+To mark text as code font, use the following:
+
+- In HTML, use the `code` element.
+- In Markdown, use backticks (`` ` ``).
+
+For information about choosing HTML or Markdown, see
+[Markdown versus HTML](#markdown).
+
+This page explains how to format code in ordinary text sentences. For more information about
+formatting and explaining placeholders, command-line syntax, and code samples, see the following
+resources:
+
+- [Formatting placeholders](#placeholders)
+- [Documenting command-line syntax](#code-syntax)
+- [Code samples](#code-samples)
+- [Code style guides](#code-samples--coding)
+- [Formatting a heading or title](https://developers.google.com/style/headings#formatting-a-heading-or-title)
+
+<a id="code-in-text--code"></a>
+
+<a id="code-in-text--some-specific-items-to-put-in-code-font"></a>
+
+#### Some specific items to put in code font
+
+The following table includes items that should be in code font, but it's not an exhaustive
+list:
+
+| Item | Recommended |
+| --- | --- |
+| Attribute names and values | The `imageURL` attribute contains the path for the image file that you can open in a browser—for example, `https://www.example.com/images/product.jpg`.  You can create a VM instance using the `e2-highcpu-16` machine type in the `us-central1-a` region. |
+| Class names | The `SnapshotDiskOperator` class includes the `generate_snapshot_name` method. |
+| Command output | The output is similar to the following:     ```         Found sysprep-specialize-script-ps1 in metadata.         ...         Finished running specialize scripts.          ``` |
+| [Command-line utility names](https://developers.google.com/style/code-in-text#tool-names), such as `gcloud`, `gsutil`, `kubectl`, and `bq` | You can use the `kubectl` tool to define a network policy. |
+| Data types | Nested data is represented as a `STRUCT` type. |
+| Database elements (such as row and column names) | The query extracts the `month`, `julianday`, and `dayofweek` values from the `datetime` and `timestamp` columns. |
+| Defined (constant) values for an element or attribute | The constant `city` has the value `"San Francisco"`. |
+| [DNS record types](https://wikipedia.org/wiki/List_of_DNS_record_types) | Create a DNS `AAAA` record in your public DNS zone that points to the IP address of the load balancer. |
+| Element names (HTML and XML) | The `script` and `df-messenger` HTML elements should be in the `body` element of your page.  A C-CDA document contains a header and a body enclosed within a `ClinicalDocument` XML element.  When you refer to an element name, don't put angle brackets (`<>`) around the element name. |
+| Enum (enumerator) names | Generated from the protobuf enum `BOOL = 1;`. |
+| Environment variable names | Set the `CHROME_REMOTE_DESKTOP_DEFAULT_DESKTOP_SIZES` environment variable to include the resolution of your monitor. |
+| Filenames, [filename extensions](#filenames--file-type-names) (if used), and paths | Open the `pg_hba.conf` file, which is typically in the `/etc/postgresql/13/main` directory. |
+| Folders and directories | The configuration information for the reader deployment is in the `opentsdb-read.yaml.tpl` file in the `deployments` folder of the guide repository. |
+| [HTTP content-type](https://www.w3.org/Protocols/rfc1341/4_Content-Type.html) values | The value of the `Content-Type` header value is required and must be set to `application/fhir+json` as defined in the FHIR specification. |
+| [HTTP status codes](#code-in-text--statuscodes) | The HTTP `500 Internal Server Error` status code indicates that the server encountered an unexpected condition that prevented it from fulfilling the request. |
+| HTTP verbs | To specify image content directly using a local image file, you can use a `POST` request. |
+| IAM role names | Grant the new service account the `roles/cloudfunctions.invoker` IAM role for the `trace` function. |
+| IP addresses | The other nodes of the cluster should contact this host on IP address `10.10.10.10.` |
+| Language keywords | The SQL statement contains the dataset table name after the `FROM` keyword in the format of `PROJECT_NAME.DATASET.TABLE_NAME`. |
+| [Method and function names](#code-in-text--methods) | The `ST_GEOPOINT` function uses the longitude and latitude of the Colosseum in Rome.  To fetch the status of the job, call the `get_job_status` method. |
+| Namespace aliases | Use Config Sync to apply the package only to the `default` namespace. |
+| [Placeholder variables](#placeholders) | Replace `SUBNETWORK_NAME` with the resource ID of the private subnet that you want the blueprint to use. |
+| Package names | The Beautiful Soup library for parsing web pages is distributed as the `beautifulsoup4` package. |
+| Port numbers | Each member Pod must have a container that's listening on TCP port `50000`. |
+| Query parameter names and values | If you want to return all contents under a directory, use the `recursive=true` query parameter with your request. |
+| Strings (such as URLs or domain names) that are used in commands and code | In IAM, a condition can specify a page that only Human Resources admins can access—for example, `https://hr.example.com`.  The `logID` field includes the domain `corpaudits.example.com`. |
+| Text input | In the **Key name** field, enter `config-management`. |
+| [UI elements](#ui-elements) that are rendered based on previously entered text (such as a server or instance name) | From the **Server name** list, select **`my-sql-cluster1`**.  Click **`my-instance`**.  If a code-formatted element appears in UI, add bold as well. For more information, see [Code in UI elements](#code-in-text--code-in-ui). |
+
+Generally, don't put quotation marks around code unless the quotation marks
+are part of the code.
+
+<a id="code-in-text--roman"></a>
+
+<a id="code-in-text--items-to-put-in-ordinary-non-code-font"></a>
+
+#### Items to put in ordinary (non-code) font
+
+The following table includes items that should not be in code font, but it's
+not an exhaustive list. If you're referring to any of these items as computer input or output,
+or as a code entity like an attribute or value, then use code font.
+
+| Item | Recommended |
+| --- | --- |
+| Domain names | The test environment is designed only for standard application offerings from example.com. |
+|<a id="code-in-text--u0023tool-names"></a> Names of products, services, and organizations | Example Organization has current and former employees who use Google products such as Google Docs and Google Sheets. |
+| URLs that the reader is supposed to follow in a browser | You can find support at https://support.example.com.  It's usually best to format a URL as a link and use descriptive link text instead of exposing the URL itself. For more information, see [Avoid URLs as link text](#cross-references--urls). |
+
+<a id="code-in-text--code-in-ui"></a>
+
+<a id="code-in-text--code-in-ui-elements"></a>
+
+#### Code in UI elements
+
+If a
+[UI element](#ui-elements--formatting)
+meets the
+[requirements for code font](#code-in-text--code),
+then use both code font and bold for that element.
+
+Recommended: In the **Network** list, select
+**`my-net-2`**.
+
+Recommended: In the **Query results** pane,
+the **`Store`** column is displayed.
+
+<a id="code-in-text--sometimes"></a>
+
+<a id="code-in-text--items-that-are-sometimes-in-code-font"></a>
+
+#### Items that are sometimes in code font
+
+The following list includes items that are sometimes in code font, but it's not an exhaustive
+list.
+
+- **Boolean values**. If you refer directly to a Boolean data type value (such
+  as `true` or `false`, or `1` or `0`), then format
+  the value as code. If you refer to the evaluation of a Boolean condition as true or
+  false, then refer to the evaluation in non-code font.
+
+  Recommended:
+
+  - If the update succeeds, returns `true`.
+  - `enableCertificateValidation`: If true, validates the SSL certificate
+    before proceeding. If false, trusts the certificate without validating it.
+- **Command-line utility names**. Often, command-line utility names are spelled the same
+  as the software project or product with which they are associated, with only differences in
+  capitalization. In such cases, use code font for the command and ordinary font for the name of
+  the project or product.
+
+  Recommended:
+
+  - Invoke the GCC 8.3 compiler using `gcc` for C programs or
+    `g++` for C++ programs.
+  - To send the file over FTP with IPv6, use `ftp -6`.
+  - The options for the `curl` command are explained on the
+    curl project website.
+  - The `apt` program includes commands from the `apt-get`
+    and `apt-cache` programs for working with APT packages.
+- **Email addresses as input or output**. If you want the reader to use the email address
+  as computer input or output, use code font. If you want the reader to treat the email address as
+  a way to contact someone or a reference to someone, use non-code font and hyperlink the email
+  address.
+
+  Recommended:
+
+  - Enter the username, not the full email address. For example, enter `alex`,
+    not `alex@example.com`.
+  - For help, contact [support@example.com](mailto:support@example.com).
+
+<a id="code-in-text--methods"></a>
+
+<a id="code-in-text--method-names"></a>
+
+#### Method names
+
+When you refer to a method name in text, omit the class name except where
+including it would prevent ambiguity.
+
+Recommended: To retrieve the zebra's
+metadata, call its `get` method.
+
+Not recommended: To retrieve the zebra's
+metadata, call its `animal.get` method.
+
+<a id="code-in-text--statuscodes"></a>
+
+<a id="code-in-text--http-status-codes"></a>
+
+#### HTTP status codes
+
+To refer to a single status code, use the following formatting and
+phrasing:
+
+an HTTP `400 Bad Request` status code
+
+In particular, call it a *status code* instead of a *response
+code* or *error code*, and put the number and the name in code font.
+If the *HTTP* is implicit from context, you can leave it out.
+
+To refer to a range of codes, use the following form:
+
+an HTTP `2xx` or `400` status code
+
+In particular, use *Nxx* (with a specific digit in place of
+N) to indicate *anything in the N00-N99
+range*, and put the status code number in code font even if you're leaving
+out the code's name.
+
+If you prefer to specify an exact range, you can do so:
+
+an HTTP status code in the `200`-`299`
+range
+
+Here, too, put the numbers in code font.
+
+<a id="code-in-text--keywords"></a>
+
+<a id="code-in-text--grammatical-treatment-of-code-elements"></a>
+
+#### Grammatical treatment of code elements
+
+In general, don't use code elements such as keywords and filenames as if they were
+English verbs or nouns. Don't inflect the name of a code element, such as to make it
+plural or possessive. Instead, include a noun after the name of the code element, and
+inflect that noun.
+
+| Recommended | Not recommended |
+| --- | --- |
+| The `ADDRESS` constant's value is defined in the `settings.h` file. | `ADDRESS`'s value is defined in `settings.h`. |
+| To add the data, send a `POST` request. | `POST` the data. |
+| To retrieve the data, send a `GET` request. | Retrieve information by `GET`ting the data. |
+| You can't close the file before opening it.  You can't call the `close` method for a file before you call `open`. | `Close`ing the file requires you to have `open`ed it first. |
+| Takes an array of extended ASCII code points (an array of `INT64` values) and returns `BYTES` values.  For `STRING` arguments, returns the original string with all alphabetic characters in uppercase. | Takes an array of extended ASCII code points (ARRAY of INT64) and returns BYTES. |
+
+<a id="code-in-text--android-linking-API-terms"></a>
+
+<a id="code-in-text--linking-api-terms-in-android"></a>
+
+#### Linking API terms in Android
+
+When you're writing code comments that you'll turn into generated reference
+documentation, link to the first instance of each element of Android APIs, such as classes, methods,
+constants, and XML attributes. Use code font and regular HTML
+`a` elements to link to this reference material.
+For later uses of the same API element in the same section, use code font
+but do not link to the reference documentation.
+
+Link `AndroidManifest.xml` elements and attributes to the API
+guide pages. Link the attribute for a particular widget or layout to its Javadoc
+in the widget or layout's API reference entry.
+
+Recommended:
+
+```
+<a href="/guide/topics/manifest/data-element.html">data</a>
+```
+
+Very common classes such as `Activity` and `Intent`
+don't need to be linked every time. If you use a term as a concept rather than a
+class, then don't put it in code font and don't capitalize it. Here are some
+objects that do not always require Javadoc links or capitalization:
+
+- activity, activities
+- service
+- fragment
+- view
+- loader
+- action bar
+- intent
+- content provider
+- broadcast receiver
+- app widget
+
+If you use one of these terms in the context of referring to an actual
+instance, use the formal class name and link to its reference page. Here are two
+examples:
+
+Recommended: The [`Activity`
+class](https://developer.android.com/reference/android/app/Activity.html) is an important part of an application's overall lifecycle...
+
+Recommended: The user interface for an
+activity is provided by a hierarchy of views—objects derived from the
+[`View` class](https://developer.android.com/reference/android/view/View.html).
+
+To link to a class or method:
+
+- To link to a class, use the class name as link text—for example:
+
+  ```
+  <a href="/reference/android/widget/TextView">TextView</a>
+  ```
+- To link to a method, use the method name as a fragment identifier. If
+  you're linking to a static method, also include the class name in the link
+  text. If you need to distinguish between overloaded versions of a particular
+  method, consider showing the full signature—for example:
+
+  ```
+  <a href="/reference/android/app/Activity.html#onCreate(android.os.Bundle)">onCreate(Bundle)</a>
+  ```
+- To link the attribute for a particular widget or layout to its Javadoc
+  in the widget or layout's API reference entry, use the URL for the page, and
+  then add the fragment identifier
+  `#attr_android:ATTRIBUTE_NAME`. For example, to link to
+  the XML attribute `android:inputType` for the `TextView`
+  widget, add the following:
+
+  ```
+  <a href="/reference/android/widget/TextView.html#attr_android:inputType>inputType</a>
+  ```
+
+---
+
+<a id="code-samples"></a>
+
+### Code samples
+
+*Source: <https://developers.google.com/style/code-samples>*
+
+This page explains how to format code samples. For more information about formatting and
+explaining code that appears in text, command-line syntax, and placeholders, see the following
+resources:
+
+- [Code in text](#code-in-text)
+- [Documenting command-line syntax](#code-syntax)
+- [Formatting placeholders](#placeholders)
+
+<a id="code-samples--guidelines"></a>
+
+<a id="code-samples--basic-guidelines"></a>
+
+#### Basic guidelines
+
+Follow these guidelines when formatting code samples:
+
+- **Follow the indentation guidelines in the relevant
+  [code style guide](#code-samples--coding)**. For most programming languages, this means using
+  spaces instead of tabs and using two spaces for each indentation level. However, some contexts
+  use four spaces for each indentation level, and some contexts use tabs. This guidance applies to
+  formatting code samples, not to [formatting commands](#code-syntax--formatting-a-command).
+- **Wrap lines** at 80 characters. If you expect readers to have a relatively narrow
+  browser window or to print out your document, consider wrapping at a smaller number of
+  characters for readability.
+- **Mark code blocks as preformatted text**. In HTML, use a `pre` element;
+  in Markdown, indent every line of the code block by four spaces.
+- **Indicate omitted code by using a comment** in the syntax of the language of your code
+  sample. Don't use three dots or the ellipsis character (`…`). If a code
+  block contains an omission, don't format the block as click-to-copy.
+
+Recommended:
+
+```
+<pre>
+function helloWorld() {
+  alert('Hello, world! This sentence is so long that it wraps onto a second
+    line.');
+}
+</pre>
+```
+
+This renders the following code block:
+
+```
+function helloWorld() {
+  alert('Hello, world! This sentence is so long that it wraps onto a second
+    line.');
+}
+```
+
+Recommended:
+
+```
+apiVersion: serving.knative.dev/v1
+kind: Service
+# Several lines of code are omitted here.
+spec:
+  template:
+    spec:
+      containers:
+      - image: IMAGE_URL
+        ports:
+        - name: h2c
+          containerPort: 8080
+```
+
+<a id="code-samples--intros"></a>
+
+<a id="code-samples--introductory-statements"></a>
+
+#### Introductory statements
+
+In most cases, precede a code sample with an introductory sentence or
+paragraph. The introduction can end with a colon or a period; usually a colon if it
+immediately precedes the sample, usually a period if there's more material (such
+as a note paragraph) between the introduction and the sample, or if the
+introduction paragraph ends in a sentence that isn't directly related to the
+sample.
+
+Recommended (ending with a period): The
+following code sample shows how to use the `get` method. For
+information about other methods, see [link]. [sample]
+
+Also recommended: The following code
+sample shows how to use the `get` method: [sample] For information about
+other methods, see [link].
+
+Not recommended (ending with a colon): The
+following code sample shows how to use the `get` method. For
+information about other methods, see [link]: [sample]
+
+For more information about how to introduce code samples, see
+[Document command-line syntax](#code-syntax).
+
+<a id="code-samples--coding"></a>
+
+<a id="code-samples--code-style-guides"></a>
+
+#### Code style guides
+
+The following public Google coding-style guides are available on GitHub:
+
+- [C++ style guide](https://google.github.io/styleguide/cppguide.html).
+- [HTML/CSS style guide](https://google.github.io/styleguide/htmlcssguide.html).
+- [Java style guide](https://google.github.io/styleguide/javaguide.html).
+- [JavaScript style guide](https://google.github.io/styleguide/javascriptguide.xml).
+- [Python style guide](https://google.github.io/styleguide/pyguide)
+- [Full list of Google's programming style guides](https://google.github.io/styleguide/)
+
+Some open source projects have their own overriding style guides. For
+example, Java code in the Android Open Source Project follows the [AOSP Java Code
+Style for Contributors](https://source.android.com/setup/contribute/code-style) guide.
+
+---
+
+<a id="code-syntax"></a>
+
+### Command-line syntax
+
+*Source: <https://developers.google.com/style/code-syntax>*
+
+<a id="code-syntax--command-line-syntax"></a>
+
+This page shows how to document command-line commands and their arguments. For more
+information about formatting code that appears in text, placeholders, and code samples, see the
+following links:
+
+- [Code in text](#code-in-text)
+- [Formatting placeholders](#placeholders)
+- [Code samples](#code-samples)
+
+<a id="code-syntax--best-practices"></a>
+
+#### Best practices
+
+When you write procedural or conceptual documentation for a command-line command, apply the
+following best practices:
+
+- **Provide an inline link to the command reference**. A good place for that link is in
+  the text that introduces the command or a series of steps.
+
+  Recommended:
+
+  To connect to the instance, use the
+  [`gcloud compute ssh` command](https://cloud.google.com/sdk/gcloud/reference/compute/ssh):
+
+  ```
+  gcloud compute ssh
+  ```
+- **Determine which arguments are needed to complete each task in the recommended way**.
+  To minimize the number of options that you need to document in non-reference content, use as
+  few optional arguments as possible. Rely on the command reference for the complete list of
+  arguments.
+- **Provide a click-to-copy command example that the reader doesn't need to edit after they
+  copy it**. If possible, include only runnable code and placeholder variables in the
+  click-to-copy example.
+
+  Some command examples contain
+  [optional arguments](#code-syntax--optional-arguments),
+  [mutually exclusive arguments](#code-syntax--set-of-two-arguments), or
+  [repeated arguments](#code-syntax--arguments-that-can-repeat)
+  that are indicated by square brackets (`[]`), pipes (`|`),
+  braces (`{}`), and ellipses (`...`). These characters can break
+  commands if they're not first removed. For that reason, avoid using these
+  arguments in click-to-copy examples.
+
+  For more information, see the
+  [Optional arguments in click-to-copy commands](#code-syntax--click-to-copy-commands)
+  section of this document.
+
+<a id="code-syntax--formatting-a-command"></a>
+
+<a id="code-syntax--format-a-command"></a>
+
+#### Format a command
+
+To mark a block of code such as a lengthy command or a code sample, use the
+following formatting:
+
+- In HTML, use the `pre` element.
+- In Markdown, use a code fence (`` \`\`\` ``).
+
+To format a command with multiple elements, do the following:
+
+- When a line exceeds 80 characters, you can safely add a line break before
+  some characters, such as a single hyphen, double hyphen, underscore, or
+  quotation marks. After the first line, indent each line by four spaces to vertically align each line
+  that follows a line break.
+- When you split a command line with a line break, each line except the
+  last line must end with the command-continuation character. Commands that don't
+  have the command-continuation character don't work.
+
+  - Linux or Cloud Shell: A backslash typically preceded with a space
+    ( `\`)
+  - Windows: A caret preceded with a space ( `^`)
+- Format placeholder text with [placeholders](#placeholders).
+- Follow the command line with a descriptive list of the placeholders
+  used in the command line. For more information, see [Explaining placeholders](#placeholders--explain-placeholders).
+- When documenting a command-line option or argument, use end puctuation for complete
+  sentences. Don't use end punctuation for single words or noun phrases, unless there is a mix of
+  sentences and noun phrases. This guidance is similar to [end punctuation in lists](#lists--capitalization-and-end-punctuation).
+  For more information, see [Google AIP guidelines for documentation](https://google.aip.dev/192#style).
+
+When you're documenting a `bash` or `sh` command, follow the
+[quotation mark style](https://google.github.io/styleguide/shellguide.html#s5.7-quoting)
+in Google's shell style guide.
+
+<a id="code-syntax--prompt"></a>
+
+#### Command prompt
+
+If your command-line instructions show multiple lines of input in one block, then start each line
+of input with the prompt symbol. If you don't want users to copy the prompt symbol when they copy
+the command, you might be able to turn off text selection for the symbol—for example, by using
+CSS.
+
+Don't show the current directory path before the prompt, even if
+part of the instruction includes changing directories. However, if the overall
+context of the command interface changes—such as from the local machine
+to a remote machine—then add an additional prompt indicator, as appropriate, for
+the new context.
+
+Recommended:
+
+Enter the following code into the terminal:
+
+```
+$ adb devices
+```
+
+The output is the following:
+
+```
+List of devices attached
+emulator-5554  device
+emulator-5556  device
+```
+
+Recommended:
+
+```
+$ adb shell
+shell@ $ screencap /sdcard/screen.png
+shell@ $ exit
+$ adb pull /sdcard/screen.png
+```
+
+When you're showing a one-line command, the command prompt
+(the `$` symbol) is optional. However, if your document includes both
+multi-line and one-line commands, then we recommend using the command prompt
+for all of the commands in the document for consistency.
+
+If your command-line instructions include a combination of input and output
+lines, we recommend using separate code blocks for input and output.
+
+Recommended:
+
+```
+$ cat ~/.ssh/my-ssh-key.pub
+```
+
+The output is similar to the following:
+
+```
+ssh-rsa KEY_VALUE USERNAME
+```
+
+<a id="code-syntax--optional-arguments"></a>
+
+#### Optional arguments
+
+Use square brackets around an argument to indicate that it's optional. If there's more than one
+optional argument, enclose each item in its own set of square brackets.
+
+Avoid using optional arguments in click-to-copy code examples. For best practices on documenting
+optional arguments with click-to-copy commands, see the
+[Best practices](#code-syntax--best-practices) and
+[Optional arguments in click-to-copy commands](#code-syntax--click-to-copy-commands)
+sections of this document.
+
+In the following example, `GROUP` is required, but
+`GLOBAL_FLAG` and `FILENAME` are optional:
+
+```
+gcloud dns GROUP [GLOBAL_FLAG] [FILENAME]
+```
+
+<a id="code-syntax--set-of-two-arguments"></a>
+
+#### Mutually exclusive arguments
+
+Use curly braces to indicate that the reader must choose one—and only one—of the
+items inside the braces. There can be more than two mutually exclusive choices. To separate each
+choice, use a pipe (`|`).
+
+Avoid using mutually exclusive arguments in click-to-copy code examples. For best practices on
+documenting mutually exclusive arguments with click-to-copy commands, see the
+[Best practices](#code-syntax--best-practices) and
+[Optional arguments in click-to-copy commands](#code-syntax--click-to-copy-commands)
+sections of this document.
+
+In the following example, choose either `FILE_1` or `FILE_2`:
+
+```
+{FILE_1|FILE_2}
+```
+
+In the following example, there are also two options:
+
+- Left side of pipe: If the source code is deployed from a cloud
+  repository, the following is required:  
+  `--source=CLOUD_SOURCE --source-url=SOURCE_URL`
+- Right side of pipe: If the source code is in a local directory:
+  - `--bucket=BUCKET` is required.
+  - `--source=LOCAL_SOURCE` is optional, as specified by the square
+    brackets.
+
+```
+{--source=CLOUD_SOURCE --source-url=SOURCE_URL | --bucket=BUCKET [--source=LOCAL_SOURCE]}
+```
+
+<a id="code-syntax--arguments-that-can-repeat"></a>
+
+#### Arguments that can repeat
+
+Use three dots and no spaces (`...`) to indicate that the reader can specify multiple
+values for the argument.
+
+Avoid using an ellipsis in click-to-copy code examples. For best practices on documenting optional
+arguments with click-to-copy commands, see the
+[Best practices](#code-syntax--best-practices) and
+[Optional arguments in click-to-copy commands](#code-syntax--click-to-copy-commands)
+sections of this document.
+
+In this example, the reader can specify multiple instances of the optional
+parameter `GLOBAL_FLAG`:
+
+```
+gcloud dns GROUP [GLOBAL_FLAG ...]
+```
+
+<a id="code-syntax--click-to-copy-commands"></a>
+
+#### Optional arguments in click-to-copy commands
+
+[Optional arguments](#code-syntax--optional-arguments),
+[mutually exclusive arguments](#code-syntax--set-of-two-arguments), and
+[repeated arguments](#code-syntax--arguments-that-can-repeat)
+contain characters (such as square brackets, curly braces, pipes, and ellipses) that can break
+commands if the reader doesn't remove them. Avoid using these types of arguments in click-to-copy
+commands. Instead, choose one of the following approaches:
+
+- **Remove the optional arguments**. As a best practice,
+  [use only the necessary arguments](#code-syntax--best-practices)
+  to complete the task for the most common use case. If possible, remove optional arguments from
+  the command; always provide a link to the command reference for the command, where readers can
+  find the full list of options. For more information, check with product management or a
+  technical support specialist for the most relevant arguments.
+
+  Recommended:
+
+  To get an aggregate list of all virtual machine (VM) instances in all zones for a project,
+  use the
+  [`gcloud compute instances list` command](https://cloud.google.com/sdk/gcloud/reference/compute/instances/list):
+
+  ```
+  gcloud compute instances list
+  ```
+
+  If you want to narrow the list of VMs to a specific zone, use the previous command with the
+  `--zones` flag.
+- **Use separate code blocks for each option**. In some cases, it might be ideal to
+  provide more than one click-to-copy code block within the same section.
+
+  Recommended:
+
+  To create a bootable Compute Engine image, use the
+  [`gcloud compute images import` command](https://cloud.google.com/sdk/gcloud/reference/compute/images/import):
+
+  ```
+  gcloud compute images import IMAGE_NAME \
+      --source-file=SOURCE_FILE
+  ```
+
+  If you're importing an image with an existing license, specify the
+  `--byol` flag:
+
+  ```
+  gcloud compute images import IMAGE_NAME \
+      --source-file=SOURCE_FILE \
+      --byol
+  ```
+- **Document optional arguments in separate tasks**. In some cases, it might be best to
+  treat different options in separate sections.
+
+  Recommended:
+
+  To create a bootable or non-bootable Compute Engine image based on an existing virtual
+  disk, use the
+  [`gcloud compute images import` command](https://cloud.google.com/sdk/gcloud/reference/compute/images/import).
+
+  <a id="code-syntax--import-a-bootable-virtual-disk"></a>
+
+  ##### Import a bootable virtual disk
+
+  If your virtual disk has a bootable operating system installed on it, run the following
+  command:
+
+  ```
+  gcloud compute images import IMAGE_NAME \
+      --source-file=SOURCE_FILE
+  ```
+
+  <a id="code-syntax--import-a-non-bootable-virtual-disk"></a>
+
+  ##### Import a non-bootable virtual disk
+
+  If your virtual disk doesn't have a bootable operating system installed on it, include the
+  `--data-disk` flag:
+
+  ```
+  gcloud compute images import IMAGE_NAME \
+      --source-file=SOURCE_FILE \
+      --data-disk
+  ```
+- **Let the reader know that the command contains optional arguments**. If you must
+  include special characters to indicate optional arguments, indicate that fact when you
+  introduce the command.
+
+  Recommended:
+
+  To create a VM with a custom name and attach one or more existing stateful disks to that VM,
+  use the
+  [`gcloud compute instance-groups managed create-instance` command](https://cloud.google.com/sdk/gcloud/reference/compute/instance-groups/managed/create-instance)
+  with one or multiple `--stateful-disk` flags. In the following example, you
+  optionally specify the `auto-delete` subflag to keep or discard each disk when the
+  VM is permanently deleted:
+
+  ```
+  gcloud compute instance-groups managed create-instance NAME \
+      --instance=VM_NAME \
+      --stateful-disk=device-name=DEVICE_NAME,source=DISK[,auto-delete=DELETE_RULE]
+  ```
+
+  For example, the following command creates a managed instance that's named
+  `db-instance` and attaches the persistent disk `db-data-disk-1` as a
+  stateful disk that is detached and preserved if its VM is deleted:
+
+  ```
+  gcloud compute instance-groups managed create-instance example-database-mig \
+      --instance=db-instance \
+      --stateful-disk=device-name=data-disk,source=projects/example-project/zones/us-east1-c/disks/db-data-disk-1,auto-delete=never
+  ```
+
+<a id="code-syntax--output"></a>
+
+#### Output from commands
+
+You don't have to show output for every command. Add output only if it adds value—for
+example, if the reader needs to copy a value from the output or if they need to verify a value
+in the output.
+
+If you are showing output, use one of the following introductory phrases to separate the command
+from the output.
+
+Recommended: The output is similar to the following:
+
+Recommended: The output is the following:
+
+If you want to explicitly call out something about the output, you can customize the introductory
+phrase.
+
+Recommended: The output is similar to the
+following, in which the `IP` column shows the IP address for each resource:
+
+To indicate that one or more lines of output are omitted from sample output, use three dots and
+no spaces (`...`) on a separate line. Do not use the ellipsis character (`…`).
+For example:
+
+```
+Reading file status
+Upload done, resetting board...
+...
+Wakeup reason: 0
+```
+
+For more information about presenting output, also see the following:
+
+- For more information about how to present output in procedures, see [Order of multiple
+  components in a step](#procedures--order-of-multiple-components-in-a-step).
+- For more information about using placeholders in output, see [Placeholders in output](#placeholders--placeholders-in-output).
+- For more information about using examples such as domain names and IP addresses in output, see [Example domains and names](#examples).
+
+<a id="code-syntax--command-line-terminology"></a>
+
+#### Command-line terminology
+
+When discussing commands and their constituent parts in the `gcloud` CLI
+and in Linux commands, follow this guidance:
+
+- Avoid mapping nomenclature of the `gcloud` CLI's commands to
+  Linux commands.
+- Linux commands can be complicated. It's wise to describe what the entire
+  command does rather than what its individual elements are called.
+- For Linux commands or commands in the `gcloud` CLI, ask yourself if the reader must
+  know the name of the command-line element or if explaining the command is sufficient.
+
+<a id="code-syntax--gcloud-commands"></a>
+
+##### gcloud commands
+
+```
+gcloud GROUP | COMMAND [--account=ACCOUNT] [--configuration=CONFIGURATION] \
+    [--flatten=[KEY,...]][--format=FORMAT] [--help] [--project=PROJECT_ID] \
+    [--quiet, -q][--verbosity=VERBOSITY; default="warning"] [--version, -v] \
+    [-h] [--log-http][--trace-token=TRACE_TOKEN] [--no-user-output-enabled]
+```
+
+For the sake of accurate classification, the `gcloud` CLI's
+syntax distinguishes between a *command* and a *command group*. In
+docs, however, command-line contents are generally referred to as commands.
+
+You can use commands (and groups) alone or with one or more flags. A
+*flag* is a Google Cloud-specific term for any element
+other than the command or group name itself. A command or flag might also
+take an *argument*, for example, a region value.
+
+<a id="code-syntax--example-command"></a>
+
+###### Example command
+
+```
+gcloud init
+```
+
+<a id="code-syntax--example-command-with-a-flag"></a>
+
+###### Example command with a flag
+
+```
+gcloud init --skip-diagnostics
+```
+
+<a id="code-syntax--example-command-with-multiple-elements"></a>
+
+###### Example command with multiple elements
+
+```
+gcloud ml-engine jobs submit training ${JOB_NAME} \
+    --package-path=trainer \
+    --module-name=trainer.task \
+    --staging-bucket=gs://${BUCKET} \
+    --job-dir=gs://${BUCKET}/${JOB_NAME} \
+    --runtime-version=1.2 \
+    --region=us-central1 \
+    --config=config/config.yaml \
+    -- \
+    --data_dir=gs://${BUCKET}/data \
+    --output_dir=gs://${BUCKET}/${JOB_NAME} \
+    --train_steps=10000
+```
+
+The preceding command consists of the following elements:
+
+- `ml-engine` is a `gcloud` command group.
+- `jobs` is an `ml-engine` command group.
+- `submit` is a `jobs` command group.
+- `training` is a `submit` command.
+- `${JOB_NAME}` is an argument that refers to an environment
+  variable called `JOB_NAME` that was set earlier.
+- `--package-path` is a flag set to a path to a Python package to build.
+- `--` in isolation separates the `gcloud` arguments that precede it from
+  the [user arguments](https://cloud.google.com/sdk/gcloud/reference/ml-engine/jobs/submit/training#USER_ARGS)
+  that follow it.
+
+In addition to the term flag, *option* is often used as a
+catchall term when you don't want to mire the reader in specialized
+nomenclature.
+
+For more information, see the
+[Cloud SDK: gcloud](https://cloud.google.com/sdk/gcloud/reference/)
+topic.
+
+<a id="code-syntax--linux-commands"></a>
+
+##### Linux commands
+
+> [!CAUTION]
+> **Caution**: Linux command syntax is notoriously complex.
+> This section covers only the most common elements. For a more detailed reference,
+> see [The Linux Command Line](http://wiki.lib.sun.ac.za/images/c/ca/TLCL-13.07.pdf).
+
+Where the `gcloud` CLI uses the catchall terms
+flag and option, Linux commands use *options*, *parameters*,
+*arguments*, and a host of specialized syntax elements. The following is an
+example:
+
+```
+find /usr/src/linux -follow -type f -name '*.[ch]' | xargs grep -iHn pcnet
+```
+
+The preceding command consists of the following elements:
+
+- `find` is the command name.
+- `/usr/src/linux` is an argument that specifies the path to look
+  in. Easier to refer to as only a path.
+- `-follow` is an option. The hyphen (`-`), often called a *dash* in
+  this context, is part of the option.
+- `-type` is an option with a value of `f`.
+- `-name` is an option with a value of `'*.[ch]'`, where
+  the asterisk (`*`) is a *metacharacter* signifying a wildcard.
+  Metacharacters are used in Linux shell commands for *globbing*, or filename
+  expansion. In addition to the asterisk, metacharacters include the question mark
+  (`?`) and caret (`^`).
+
+The results of the first command are redirected by using a *pipe*
+(`|`) to the `xargs grep -iHn pcnet` command. Other
+redirection symbols include the greater than symbol (`>`), less than symbol
+(`<`), left double angle quotation mark (`<<`), and right double
+angle quotation mark (`>>`). Redirection means capturing
+output from a file, command, program, script, or even code block within a script
+and sending it as input to another file, command, program, or script.
+
+<a id="code-syntax--linux-signals"></a>
+
+##### Linux signals
+
+Linux signals require vocabulary choices that
+are generally discouraged elsewhere in documentation. We recommend using the terms in the
+following table *only* in the context of process control:
+
+| Signal | Description |
+| --- | --- |
+| `SIGKILL` | Signal sent to *kill* a specified process, all members of a specified process group, or all processes on the system. `SIGKILL` cannot be caught, blocked, or ignored. Do not substitute *cancel*, *end*, *exit*, *quit*, *stop*, or *terminate*. |
+| `SIGTERM` | Signal sent as a request to *terminate* a process. Although similar to `SIGKILL`, this signal gives the process a chance to clean up any child processes that might be running. Do not substitute *cancel*, *end*, *exit*, *quit*, or *stop*. |
+| `SIGQUIT` | Signal sent from a keyboard to *quit* a process. Some processes can catch, block, or ignore a quit signal. Do not substitute *cancel*, *end*, *exit*, *quit*, or *stop*. |
+| `SIGINT` | Signal sent to *interrupt* a process immediately. The default action of this signal is to terminate a process gracefully. It can be handled, ignored, or caught. It can be sent from a terminal—for example, when a user presses `Control+C`. Do not substitute *suspend*, *end*, *exit*, *pause*, or *terminate*. |
+| `SIGPAUSE` | Signal that tells a process to *pause*, or *sleep*, until any signal is delivered that either terminates the process or invokes a signal-catching function. Do not substitute *cancel* or *interrupt*. |
+| `SIGSUSPEND` | Signal sent to temporarily *suspend* execution of a process. Used to prevent delivery of a particular signal during the execution of a critical code section. Do not substitute *pause* or *exit*. |
+| `SIGSTOP` | Signal sent to *stop* execution of a process for later continuation (upon receiving a `SIGCONT` signal). `SIGSTOP` cannot be caught, blocked, or ignored. Do not substitute *cancel*, *end*, *exit*, *interrupt*, *quit*, or *terminate*. |
+
+---
+
+<a id="placeholders"></a>
+
+### Placeholder formatting
+
+*Source: <https://developers.google.com/style/placeholders>*
+
+This page explains how to format placeholders in commands, code samples, and text
+strings. This page doesn't explain how to implement visual styling for placeholders, but it does
+show examples of how Google developer documentation style renders placeholders as visually
+distinct from other text.
+
+For more information about formatting code, command-line syntax, and code samples, see the
+following links:
+
+- [Code in text](#code-in-text)
+- [Documenting command-line syntax](#code-syntax)
+- [Code samples](#code-samples)
+
+Placeholders in sample code and commands represent values that the reader must replace when they use
+the sample input. Placeholders in example output can also represent other values that vary. In
+general, a placeholder has a descriptive name as a default value.
+
+For example, the placeholder `PROJECT_ID` represents a project ID in sample
+code, commands, and example output.
+
+In example output, the placeholder `HTTP_RESPONSE_CODE` represents an
+HTTP response code; the reader isn't expected to set this to a specific value.
+
+<a id="placeholders--placeholder-variables"></a>
+
+<a id="placeholders--placeholders"></a>
+
+#### Placeholders
+
+When you create placeholders follow this general guidance around using the letter
+*x*:
+
+- In general, don't use a single *x* or a series of *x*'s as placeholders; use a more
+  informative placeholder.
+- In some contexts (such as HTTP status codes), a series of *x*'s is the standard,
+  so it's OK to use (for example) *xx* in those cases.
+
+There are several ways to format placeholders, depending on whether you're
+working in HTML or Markdown, or whether the placeholder is inline, in a code block, or in a
+paragraph. For details, see the following sections.
+
+<a id="placeholders--placeholders-in-inline-text"></a>
+
+##### Placeholders in inline text
+
+If your sample code and command placeholders occur in a sentence, use the following formatting:
+
+- In HTML, wrap variable placeholders by using the `var`
+  element, like this:
+
+  ```
+  <code><var>PLACEHOLDER_NAME</var></code>
+  ```
+- In Markdown, wrap inline placeholders in backticks (`), and use an
+  asterisk (*) before the first backtick and after the second one
+  (`` *`PLACEHOLDER_NAME`* ``).
+
+If your placeholder does not represent a code sample or command, use the following formatting:
+
+- In HTML, wrap placeholders by using the `var`
+  element, like this:
+
+  ```
+  <var>PLACEHOLDER_NAME</var>
+  ```
+
+<a id="placeholders--placeholders-in-code-blocks"></a>
+
+##### Placeholders in code blocks
+
+If your placeholders are in a block of code, use the following formatting:
+
+- In HTML, wrap the code block in a `pre` element,
+  and tag placeholders with `var` elements:
+
+  ```
+        <pre>
+        gcloud compute forwarding-rules create <var>FORWARDING_RULE_NAME</var> \
+            --global | --region=<var>REGION</var> \
+            --load-balancing-scheme=<var>LOAD_BALANCING_SCHEME</var> \
+            --network=<var>NETWORK</var> \
+            ...
+        </pre>
+        
+  ```
+- In Markdown, wrap the code block in a code fence (\`\`\`). Inside a
+  code fence, you can't apply formatting like bold or italic.
+
+  ```
+  ```
+  PLACEHOLDER_NAME
+  ```
+  ```
+
+<a id="placeholders--placeholder-text"></a>
+
+##### Placeholder text
+
+**Use uppercase characters with underscore delimiters.**
+
+For example, in HTML:
+
+Recommended:
+
+- `.../<var>API_NAME</var>`
+- `.../<var>METHOD_NAME</var>`
+
+Not recommended:
+
+- `.../<var>API-name</var>`
+- `.../<var>API_name</var>`
+- `.../<var>API name</var>`
+- `.../<var>api_name</var>`
+- `.../<var>api-name</var>`
+- `.../<var>apiName</var>`
+
+In Markdown:
+
+Recommended:
+
+- `.../*API_NAME*`
+- `.../*METHOD_NAME*`
+
+If the context in which your placeholders appear makes using
+uppercase characters with underscore delimiters a bad idea, use something else
+that makes sense to you, but be internally consistent.
+
+**Don't include possessive adjectives in placeholders.**
+
+Not recommended:
+
+- `.../<var>MY_API_NAME</var>`
+- `.../<var>YOUR_API_NAME</var>`
+
+> [!NOTE]
+> **Note**: You can mark up command-line syntax with [brackets](#code-syntax--optional-arguments), [braces](#code-syntax--set-of-two-arguments), and [ellipses](#code-syntax--arguments-that-can-repeat). Don't put the
+> brackets, braces, or ellipses in the `var` element.
+
+<a id="placeholders--explain-placeholders"></a>
+
+#### Explain placeholders
+
+When you use a placeholder in text or code, explain the placeholder the first time you use it.
+It's not necessary to repeat the explanation in the document unless doing so might benefit the
+reader—for example, in circumstances such as the following:
+
+- Your document is lengthy.
+- You've introduced several other placeholders in a long procedure.
+- Your document isn't intended to be read from beginning to end.
+
+The following is an example of a command that uses a placeholder with an explanation of that
+placeholder:
+
+```
+<pre class="devsite-click-to-copy">
+gcloud compute instances create <var>INSTANCE_NAME</var> \
+    --metadata enable-guest-attributes=TRUE
+</pre>
+
+<p>Replace <code><var>INSTANCE_NAME</var></code> with the name that
+you want your new VM instance to have.</p>
+```
+
+<a id="placeholders--single-placeholder"></a>
+
+##### Single placeholder
+
+Use the following format for a single placeholder:
+
+- Replace PLACEHOLDER with a description of what
+  the placeholder represents.
+
+Recommended:
+
+1. Stream the build logs to the Google Cloud console:
+
+   ```
+   gcloud builds log --stream=BUILD_ID
+   ```
+
+   Replace `BUILD_ID` with the ID of the `WORKING` build that
+   you copied in the preceding step.
+
+<a id="placeholders--two-or-more-placeholders"></a>
+
+##### Two or more placeholders
+
+Use the following format for two or more placeholders:
+
+- Follow the command line with a descriptive list of the placeholders
+  used in the command line. Explain what each placeholder represents
+  even if the placeholder value is intuitive to you.
+- Introduce this list with *Replace the following:*
+- List the placeholders in the order in which they appear in the command line.
+- Tag each placeholder in a code sample or command with `code` and
+  `var` elements, followed by a
+  [colon and a description that starts with a lowercase letter](#colons).
+  For
+  non-code samples, remove the `code` elements—for example:
+
+  ```
+  <li><code><var>INSTANCE_NAME</var></code>: description</li>
+  ```
+- If the description contains an example, introduce it with an *em dash* or
+  *such as*—for example:
+
+  ```
+  <li><code><var>INSTANCE_NAME</var></code>: description&mdash;for example,...</li>
+  ```
+
+  ```
+  <li><code><var>INSTANCE_NAME</var></code>: description, such as...</li>
+  ```
+- Each item in the list follows our [list style](#lists).
+
+Recommended:
+
+1. Set the maximum concurrency target for a new reservation:
+
+   ```
+       bq mk \
+           --project_id=ADMIN_PROJECT_ID \
+           --location=LOCATION \
+           --target_job_concurrency=CONCURRENCY \
+           --reservation \
+           RESERVATION_NAME
+   ```
+
+   Replace the following:
+
+   - `ADMIN_PROJECT_ID`: the project that owns the reservation
+   - `LOCATION`: the location of the reservation
+   - `CONCURRENCY`: the maximum concurrency target
+   - `RESERVATION_NAME`: the name of the reservation
+
+Recommended:
+
+1. In Cloud Shell, set the environment variables:
+
+   ```
+   export ONPREM_PROJECT=ON_PREM_PROJECT_NAME \
+       export ONPREM_ZONE=ZONE
+   ```
+
+   Replace the following:
+
+   - `ON_PREM_PROJECT_NAME`: the Google Cloud project
+     name for your on-premises project. You can find your project number on the
+     [Dashboard](https://console.cloud.google.com/home/dashboard)
+     page of the Google Cloud console.
+   - `ZONE`: a [Google Cloud
+     zone](https://developers.google.com/compute/docs/regions-zones#identifying_a_region_or_zone) that's close to your location—for example, `us-east1`.
+
+<a id="placeholders--placeholders-in-output"></a>
+
+##### Placeholders in output
+
+If you provide a code output example, explain any placeholders that appear in
+sample output:
+
+- Use `var` elements to identify the placeholder text in
+  the output.
+- Follow the example output with a list of the placeholders used in the
+  example.
+- Introduce the list of placeholders with *This output includes the
+  following values:*
+- List the placeholders in the order in which they appear in the
+  example.
+- Tag each placeholder with a `var` element,
+  followed by a colon and a description that starts with a lowercase letter—for example:
+
+  ```
+  <li><code><var>INSTANCE_NAME</var></code>: description</li>
+  ```
+- If the description contains an example, introduce it with an *em dash* or
+  *such as*—for example:
+
+  ```
+  <li><code><var>INSTANCE_NAME</var></code>: description&mdash;for example,...</li>
+  ```
+
+  ```
+  <li><code><var>INSTANCE_NAME</var></code>: description, such as...</li>
+  ```
+
+For more information, see [Output from commands](#code-syntax--output).
+
+Recommended:
+
+<a id="placeholders--response"></a>
+
+###### Response
+
+The output is similar to the following:
+
+```
+{
+ "name": "operations/build/PROJECT_ID/OPERATION_ID",
+ "metadata": {
+  "@type": "type.googleapis.com/google.devtools.cloudbuild.v1.BuildOperationMetadata",
+  "build": {
+   "id": "BUILD_ID",
+   "status": "QUEUED",
+   "createTime": "2019-09-20T15:55:29.353258929Z",
+   "steps": [
+    {
+     "name": "gcr.io/compute-image-tools/gce_vm_image_import:release",
+     "env": [
+      "BUILD_ID=BUILD_ID"
+     ],
+     "args": [
+      "-timeout=7056s",
+      "-image_name=IMAGE_NAME",
+      "-client_id=api",
+      "-data-disk",
+      "-source_file=SOURCE_FILE"
+     ]
+    }
+   ],
+   "timeout": "7200s",
+   "projectId": "PROJECT_ID",
+   "logsBucket": "gs://PROJECT_NUMBER.cloudbuild-logs.googleusercontent.com",
+   "options": {
+    "logging": "LEGACY"
+   },
+   "logUrl": "https://console.cloud.google.com/gcr/builds/BUILD_ID?project=PROJECT_NUMBER"
+  }
+ }
+}
+```
+
+This output includes the following values:
+
+- `PROJECT_ID`: the project ID for the project that
+  the image was imported into
+- `OPERATION_ID`: the ID of the import operation
+- `BUILD_ID`: the ID of the build for the import
+  operation
+- `IMAGE_NAME`: the name of the image to be
+  imported
+- `SOURCE_FILE`: the URI for the image in Cloud
+  Storage—for example, `gs://my-bucket/my-image.vmdk`
+- `PROJECT_NUMBER`: the number for the import
+  project
+
+---
+
+<a id="ui-elements"></a>
+
+### UI elements and interaction
+
+*Source: <https://developers.google.com/style/ui-elements>*
+
+<a id="ui-elements--task-focus"></a>
+
+#### Focus on the task
+
+When practical, state instructions in terms of what the reader
+should accomplish, rather than focusing on the widgets and gestures.
+By avoiding reference to UI elements, you help the reader understand
+the purpose of an instruction, and it can help future-proof
+procedures.
+
+Recommended: Refresh the page.
+
+Recommended: Expand the **Advanced options**
+section.
+
+However, know the audience and understand the context. In some cases, the point
+of a procedure is to guide the reader through elements on the page. Or the UI might not be obvious,
+and it's helpful to explain the gestures for completing a step. Provide the level of detail
+that seems useful for the intended audience.
+
+Recommended: Click **Refresh**.
+
+Recommended: To expand the **Advanced
+options** section, click the
+expander arrow.
+
+The rest of this page focuses on scenarios where you've decided it's
+useful to explicitly discuss UI elements.
+
+For information about writing procedures, see [Procedures](#procedures).
+
+<a id="ui-elements--formatting"></a>
+
+#### Format names of UI elements
+
+When referring to any UI element by name, put its name in bold, using the
+`b` element in HTML or `**` in Markdown. This
+includes names for buttons, menus, dialogs, windows, list items, or any other
+feature on the page that has a visible name. Don't use code font for UI elements,
+unless it's an element that meets the [requirements for code font](#code-in-text).
+In that case, use both code font and bold.
+
+> [!NOTE]
+> **Note**: The reason for using the [`b`
+> element](https://html.spec.whatwg.org/multipage/semantics.html#the-b-element) is that in modern HTML, `b` connotes text to which
+> you want to draw visual attention, whereas the [`strong`
+> element](https://html.spec.whatwg.org/multipage/semantics.html#the-strong-element) indicates strong importance.
+
+Don't make an official feature name or product name bold, except when it
+directly refers to an element on the page that uses the name (such as a window
+title or button name).
+
+Recommended:
+In the **New project** window, select the **New activity**
+checkbox, and then click **Next**.
+
+Not recommended:
+In the New Project window, select "New Activity", and then click the
+"Next" button.
+
+If you document a UI element outside the context of a procedure, try to provide context for the
+element.
+
+Recommended: The service lets you check the status of all
+jobs in the **Current jobs** section of the service console.
+
+Not recommended: The service lets you check the status of
+all jobs in the **Current jobs** section.
+
+<a id="ui-elements--capitalization"></a>
+
+#### Use appropriate capitalization
+
+In most cases, follow the capitalization as it appears on the page. However,
+if labels are inconsistent or they're all uppercase, use sentence case.
+
+| Guidance | Recommended | Not recommended |
+| --- | --- | --- |
+| When a label is all uppercase, use sentence case. | Click **Refresh**. | Click **REFRESH**. |
+| When referring to multiple labels that are inconsistently cased, use sentence case for all of the labels. | Click **New project**, and then click **New activity**. | Click **NEW PROJECT**, and then click **New Activity**. |
+
+<a id="ui-elements--referring-to-ui"></a>
+
+<a id="ui-elements--refer-to-ui-elements"></a>
+
+#### Refer to UI elements
+
+Don't use UI elements as if they were English verbs or nouns.
+
+| Recommended | Not recommended |
+| --- | --- |
+| In the **Name** field, enter an account name. | **Name** the account. |
+| To save the settings, click **Save**. | **Save** the settings. |
+| In the **Service account ID** field, enter a name.  For **Service account ID**, enter a name. | Specify a **Service account ID**. |
+
+<a id="ui-elements--ui-terms"></a>
+
+#### Terminology and usage
+
+A user interface can contain a variety of UI elements. In general, focus on
+the feature and its functionality, not the UI element. If you think it adds
+clarity for the reader, use the name of the UI element. For example, both of the
+following sentences are valid:
+
+Recommended:
+Go to **File > Tools**.
+
+Recommended:
+In the **File** menu, click **Tools**.
+
+Don't use slang terms for UI elements—for example,
+*hamburger icon* or *zippy*. For more information, see
+[Buttons and icons](#ui-elements--buttons).
+
+Recommended:
+To expand the **Advanced options** section, click the
+expander arrow.
+
+Recommended:
+Expand **Advanced options**.
+
+Not recommended:
+To expand the **Advanced options** section, click the zippy.
+
+The following sections define some terms to use when referring to UI
+elements.
+
+For prepositions to use with these elements, see the
+[Prepositions](#ui-elements--prepositions) table.
+
+<a id="ui-elements--term-windows"></a>
+
+##### Windows, pages, dialogs, panes, and sections
+
+Most often, a *window* is the entire application window in a desktop
+environment. However, it can also refer to modular application elements that you
+can open and close. For example, in Android Studio, several windows are
+available in the **View > Tool Windows** menu.
+
+Recommended:
+In the **MyApp** window, click **Edit**.
+
+Not recommended:
+In the **MyApp** page, click **Edit**.
+
+*Page* is the preferred term when referring to a web page in general and to a subpage
+of a console in particular. For more information, see
+[console](#word-list--console).
+
+Recommended:
+In the Google Cloud console, go to the **Deployments** page.
+
+Not recommended:
+In the Google Cloud console, go to the **Deployments** window.
+
+A *dialog* is a smaller window that is usually detached from the
+main application window and appears in front of the window.
+
+Recommended:
+In the **Welcome** dialog, click **OK**.
+
+Not recommended:
+In the **Welcome** pop-up window, click **OK**.
+
+A *pane* (or *panel*) is typically a distinct rectangular region within a larger
+browser or application window. A pane or panel can often be tightly coupled to surrounding UI
+regions, whereas a window is distinctly separate and can be hidden. Do not use terms such as
+*window*, *section*, *area*, or *column* to refer to a pane or panel.
+
+Recommended:
+In the **Create service account** pane, click **New**.
+
+Not recommended:
+In the **Create service account** section, click **New**.
+
+A *section* is a labeled grouping of options and controls, usually within a window, pane, or
+panel. Do not use terms such as *area* or *column* to refer to a section.
+
+Recommended: In the **Create metric** pane, do the
+following:
+
+- In the **Metric type** section, select **Counter**.
+- In the **Labels** section, click **Add label**.
+
+<a id="ui-elements--term-menus"></a>
+
+##### Menus and menu bars
+
+In a desktop application, the *menu bar* appears at the top of the
+window or at the top of the screen; it's a set of *menus* (such as
+**File** or **Edit**), each of which is a set of related
+*commands* and/or nested submenus.
+
+To refer to an item in a menu, use the term *command*, not *choice*, *menu item*,
+or *option*. Exception: if you're documenting how to build an interface,
+you can use *menu item*.
+
+To refer to a menu, use the form *the **LABEL_NAME** menu.*
+
+To tell the reader where to find a command in a menu or submenu, use a phrase like
+*In the **File** menu, select **Open**.*
+
+Don't use *drop-down* as a synonym for *menu*. See
+[drop-down](#word-list--drop-down).
+
+<a id="ui-elements--angle-brackets"></a>
+
+###### Use angle brackets
+
+Another option is to use angle brackets (>). If you use angle brackets, follow these
+guidelines:
+
+- Put a nonbreaking space (`&nbsp;`) before each angle bracket.
+- Don't bold each menu name separately; instead, enclose the entire sequence in a single bold
+  tag (`<b>...</b>` or `**...**`).
+- Wrap the angle bracket with a span tag and add an `aria-label` attribute with
+  *and then* text
+  (for example, `<span aria-label="and then">></span>`).
+  Otherwise, some screen readers might read `>` as "greater than."
+
+In the following example, the text renders as *Select **View >
+Tools > Developer Tools***. A screen reader
+interprets this as *Select View and then Tools and then Developer Tools*.
+
+<a id="ui-elements--html"></a>
+
+##### HTML
+
+```
+Select <b>View&nbsp;<span aria-label="and then">></span> Tools&nbsp;<span aria-label="and then">></span> Developer Tools</b>.
+```
+
+<a id="ui-elements--markdown"></a>
+
+##### Markdown
+
+```
+Select **View&nbsp;<span aria-label="and then">></span> Tools <span aria-label="and then">></span> Developer Tools**.
+```
+
+This notation is useful for abbreviating a longer phrase like *In the
+**File** menu, select **Open**.* However, this notation applies only to
+menu items. Don't use it to describe a combination of different UI elements.
+
+Recommended: Select
+**MyApp > Preferences**, and then select the
+**Languages** preference pane.
+
+Not recommended:
+Select **MyApp** > **Preferences** > **Languages** >
+**+** > **CSS**.
+
+<a id="ui-elements--term-navigation-menu"></a>
+
+##### Navigation menu
+
+A *navigation menu* is a control—usually a pane or window—that contains a list of items
+that the user can click to go to pages in an application or website. Don't use the terms
+*navigation bar*, *navigation pane*, *navigation panel*, or
+*navigation window* for such a control.
+
+Recommended:
+In the BigQuery navigation menu, click **Scheduled queries**.
+
+<a id="ui-elements--term-toolbars"></a>
+
+##### Toolbar
+
+A *toolbar* is a set of buttons for common user actions. A toolbar
+button that includes a menu is called a *menu button*. Refer to the
+toolbar by name if you think that the user needs help finding a button.
+
+Recommended:
+On the Google Cloud console toolbar, click
+**Notifications**.
+
+Recommended: Click
+**Notifications**.
+
+<a id="ui-elements--buttons"></a>
+
+##### Buttons and icons
+
+A *button* initiates an action when clicked (or tapped, in the case
+of a touchscreen). To refer to a button, use the button's label.
+
+Recommended: Click **OK**.
+
+Not recommended: Click the "OK"
+button.
+
+An icon is a symbol or image that represents an object or a function. An icon
+can be a button as well. If the button includes an icon, write the name of the
+button as shown in the tooltip, and add the button icon before the
+name. If you need to use a space between the icon and the name for readability,
+use a nonbreaking space.
+
+Recommended: Click
+**Settings and utilities**.
+
+Not recommended: Click
+.
+
+If the icon tooltip is identical to the name of the icon, use an
+[empty `alt` attribute](#images--alt-text).
+
+If you're unsure of the name of the icon, inspect the element using browser
+tools. In many cases, a visual element like an icon has an ARIA attribute
+that provides a textual description of the element for use by screen readers.
+To inspect an element, right-click the element and select
+**Inspect** or **Inspect element**, depending on your browser. Look for one of the following
+types of labels: `aria-labelledby`, `aria-label`,
+`aria-describedby`, `label`, `placeholder`, or `title`.
+For more information, see
+[Using aria-label](https://www.w3.org/TR/WCAG20-TECHS/ARIA14.html)
+and
+[Accessible Name and Description calculation](https://www.w3.org/TR/html-aapi/#accessible-name-and-description-calculation).
+
+If a button with an icon doesn't include a tooltip, submit a bug report
+requesting that a tooltip be added. Tooltips are crucial for accessibility, and
+for documentation and discoverability in general.
+
+Recommended: Click ![](https://developers.google.com/static/style/images/icon-add.png) **Add**.
+
+Not recommended: Click the ![hammer icon](https://developers.google.com/static/style/images/icon-add.png) icon.
+
+If a UI element name ends with an ellipsis (...), leave out the ellipsis.
+
+Recommended: Click **Browse**.
+
+Not recommended: Click
+**Browse ...**.
+
+Don't use directional language to orient the reader, such as *above*,
+*below*, or *right-hand side*. Phrases like those don't work well for
+accessibility or for localization. If a UI element is hard to find, provide a
+screenshot.
+
+Recommended: Click **Menu**.
+
+Not recommended: In the left-side panel,
+click the button with three lines.
+
+<a id="ui-elements--difficult-to-find-ui-elements"></a>
+
+###### Difficult-to-find UI elements
+
+If you have UI elements that are difficult to find, consider one of the following options
+as an alternative to using directional language, which can be problematic for
+accessibility and localization reasons.
+
+- Use the button icon along with its name as shown in the button tooltip.
+
+  Recommended: Click
+  **Refresh**.
+- Add context to help the user find the element.
+
+  Recommended: On the Cloud Run toolbar, click
+  **Refresh**.
+- Use a screenshot.
+
+  Recommended: In the list of services, click
+  **Column display options**.
+
+  ![List of services.](https://developers.google.com/static/style/images/list-of-services.png)
+
+  For more information about when and how to use screenshots, see
+  [Diagrams, figures, and other images](#images).
+
+<a id="ui-elements--term-tab"></a>
+
+##### Tab
+
+A *tab* is a navigation element that looks like a file tab. To refer
+to a tab, use the form *the **LABEL_NAME** tab*.
+
+Recommended: Select
+**Tools > Options**, and then click the **Edit**
+tab.
+
+<a id="ui-elements--term-textbox"></a>
+
+##### Text box
+
+A *text box* is a box that the user can type in. Use
+*box* and the form *the **LABEL_NAME** box*. Format the text
+that the user types by using the `code` element in HTML, or by using code
+formatting (monospace) in other markup.
+
+Recommended: In the **Owner** box,
+enter your name.
+
+Recommended: In the **Name** box,
+enter `wsfc-1`.
+
+In Google Cloud, use
+*field* instead of *box*.
+
+In Google Workspace documentation, use
+*field* instead of *box*.
+
+Recommended: In the **Instance** field,
+specify a value less than 64 characters long.
+
+<a id="ui-elements--term-listbox"></a>
+
+##### List box, combo box, and spin box
+
+A *list box* is a box that offers the user a list of items. To refer to a list box, use
+the form *the **LABEL_NAME** list* or
+*the **LABEL_NAME** box*, whichever is clearer.
+
+Recommended: In the **Item** list, select
+**Desktop**.
+
+A *combo box* is a combination of a text box and a list box. To
+refer to a combo box, use the form *the **LABEL_NAME** box*. To refer to
+entering a value into a combo box, use the verbs *type or select* or *enter*.
+
+Recommended: In the **Font** box, type
+or select the font that you want to use.
+
+A *spin box* is a box that lets the user choose a value by clicking
+arrows or by typing. To refer to a spin box, use the form *the
+**LABEL_NAME** box*. To refer to entering a value into a spin
+box, use the verb *enter*.
+
+Recommended: In the **Font Size** box,
+enter a font size.
+
+<a id="ui-elements--term-checkbox"></a>
+
+##### Checkbox
+
+A *checkbox* is a small box that indicates whether an option is
+on or off. To refer to a checkbox, use the form *the
+**LABEL_NAME** checkbox*.
+
+Be wary of using the verbs *check* and *uncheck*, which can be ambiguous; it's often
+best to use *select* and *clear* instead.
+
+Recommended: Select the **Automatically
+check for updates** checkbox.
+
+Recommended: Clear the
+**Bookmarks** checkbox.
+
+If you need to refer to the state of the checkbox, it's often best to refer to it as
+*selected* or *not selected*.
+
+Recommended: Make sure that the
+**Bookmarks** checkbox is selected.
+
+Recommended: Make sure that the
+**Bookmarks** checkbox isn't selected.
+
+<a id="ui-elements--term-radiobutton"></a>
+
+##### Radio button
+
+A *radio button* is a small button used to choose one item from a
+group of mutually exclusive options. To refer to a radio button, use the radio
+button's label, or refer to the group of buttons by its label.
+
+Recommended: Select **Do not remember
+passwords**.
+
+Recommended: For **Startup mode**,
+select an option.
+
+<a id="ui-elements--term-expander"></a>
+
+##### Expander arrow
+
+An *expander arrow* is the UI element used to expand or collapse a section of
+navigation or content. Avoid referring to these explicitly in documentation, but when you do, use
+the terms *expander arrow* and *expandable section* rather than terms like
+*expando* or *zippy*.
+
+Recommended: To expand the
+**Advanced options** section, click the
+expander arrow.
+
+Not recommended: To expand
+the **Advanced options** section, click the zippy.
+
+<a id="ui-elements--term-toggle"></a>
+
+##### Toggle
+
+A *toggle* is the UI element that switches back and forth between on and off
+states. Don't use the word *toggle* as a verb. Describe the action that you want the
+user to take.
+
+Recommended: To turn on the setting, click
+the **Wi-Fi** toggle.
+
+In some cases, you might not know what state the toggle is in before the user interacts with it
+so be clear what position the toggle should be in.
+
+Recommended: In **Settings**, click
+the **Magic mode** toggle to the on position.
+
+<a id="ui-elements--keyboard"></a>
+
+#### Press and type keyboard keys
+
+To indicate that the user should press a given keyboard key or
+combination, use the `kbd` element.
+
+The following is an example of a `<kbd>` tag:
+
+Recommended:
+`Press <kbd>Control+C</kbd>.`
+
+When rendered, the text appears as follows:
+
+Recommended: Press `Control+C`.
+
+If you're working with non-HTML markup, use monospace formatting, which is how the
+`kbd` element renders.
+
+To refer to a letter key, use uppercase instead of lowercase.
+
+Recommended: To save, press
+`Control+S`.
+
+Not recommended: To save, press
+`Control+s`.
+
+To refer to a key that the user types to enter that key's value as text input,
+use the `code` element, not the `kbd` element.
+For more information, see [Code font](#text-formatting--code-font).
+
+To refer to a keyboard key, use the key's name. If that's ambiguous, use the
+form *the `KEY_NAME` key*.
+
+Recommended: Press `Esc`.
+
+Recommended: Press the `Esc` key.
+
+Spell out the names of modifier keys such as Command, Control, Option, and
+Shift. Don't use symbols for those keys. To refer to a key combination, use the
+form *`MODIFIER+KEY_NAME`*.
+
+Recommended: Press
+`Control+V`.
+
+When you provide shortcuts for multiple operating systems, put the macOS shortcut in
+parentheses after the Windows and Linux shortcut.
+
+Recommended: To copy, press
+`Control+C` (or `Command+C` on macOS).
+
+Not recommended: To copy, press
+`Ctrl+C` (`⌘+C`).
+
+To refer to a key or combination that uses the Shift key, use the form
+*`MODIFIER+Shift+KEY_NAME`*.
+
+Recommended: Press
+`Control+Shift+?`.
+
+Spell out the names of characters that could be confusing in a keyboard
+shortcut, such as comma, hyphen, period, and plus.
+
+To refer to a keyboard shortcut, use either *keyboard shortcut* or *key
+combination*.
+
+To refer to pressing a key or combination to cause an action to occur, use
+the verb *press*. To refer to typing a key or combination as part of text, use
+the verbs *enter* or *type*.
+
+<a id="ui-elements--prepositions"></a>
+
+#### Prepositions
+
+When documenting the UI, use the following prepositions.
+
+| Preposition | UI element | Recommended |
+| --- | --- | --- |
+| in | dialogs  fields  lists  menus  panes  windows | In the **Alert** dialog, click **OK**.  In the **Name** field, enter `wsfc-1`.  In the **Item** list, select **Desktop**.  In the **File** menu, click **Tools**.  In the **Metrics** pane, click **New**.  In the **Task** window, click **Start**. |
+| on | pages  tabs  toolbars | On the **Create an instance** page, click **Add**.  On the **Edit** tab, click **Save**.  On the **Dashboard** toolbar, click **Edit**. |
+
+<a id="ui-elements--verbs"></a>
+
+#### Verbs in procedures
+
+To describe an action on the page, use the following verbs. For more
+information about each verb, see its corresponding entry on the
+[word list](#word-list).
+
+- [Click](#word-list--click)
+- [Choose](#word-list--choose)
+- [Drag](#word-list--drag)
+- [Enable](#word-list--enable)
+- [Enter, type](#word-list--enter)
+- Go to (see [scroll](#word-list--scroll))
+- [Hold the pointer over](#word-list--hold-the-pointer-over)
+- [Press](#word-list--press)
+- [Select](#word-list--select)
+- [Tap](#word-list--tap)
+- [Turn on, turn off](#word-list--turn-on)
+
+For information about writing procedures, see [Procedures](#procedures).
+
+---
+
+## HTML and CSS
+
+<a id="semantic-tagging"></a>
+
+### HTML and semantic tagging
+
+*Source: <https://developers.google.com/style/semantic-tagging>*
+
+<a id="semantic-tagging--intro"></a>
+
+Use HTML elements for the purposes that they were designed for. For example, when
+you give the title of a standalone work (such as a book or a movie), mark it
+with a [`cite`
+element](https://html.spec.whatwg.org/multipage/text-level-semantics.html#the-cite-element). For more information about semantic tagging, see [Semantics in HTML](https://developer.mozilla.org/en-US/docs/Glossary/Semantics#Semantics_in_HTML)
+on the MDN web documents site.
+
+In situations where there are no semantically relevant HTML elements, use CSS
+or the few HTML elements that convey visual style without semantics.
+
+<a id="semantic-tagging--visual-formatting"></a>
+
+#### Visual formatting
+
+If you want to achieve specific visual results, don't use HTML elements that
+convey different semantics.
+
+In particular, follow these guidelines:
+
+- Don't use frames or tables for layout; instead, use your site's CSS to lay out the page.
+- Don't use the heading elements (such as `h1` and
+  `h2`) to visually style text; instead, use those elements
+  only for hierarchically structured headings, and use CSS for visual style.
+- The [`em`
+  element](https://html.spec.whatwg.org/multipage/text-level-semantics.html#the-em-element) indicates emphasis, not italics as such. Don't use it to italicize
+  something that isn't meant to be emphasized; instead, use the [`i`
+  element](https://html.spec.whatwg.org/multipage/text-level-semantics.html#the-i-element) for non-emphasis italics.
+- The [`strong`
+  element](https://html.spec.whatwg.org/multipage/semantics.html#the-strong-element) indicates strong importance, not bold as such. To bold a word that
+  doesn't merit strong importance, use the [`b`
+  element](https://html.spec.whatwg.org/multipage/text-level-semantics.html#the-b-element).
+- The [`br`
+  element](https://html.spec.whatwg.org/multipage/text-level-semantics.html#the-br-element) is intended "only for line breaks that are actually part of the content,
+  as in poems or addresses." Don't use it to adjust the spacing between lines.
+  Instead, use elements like `p` to semantically mark the
+  text, and use CSS to adjust line spacing.
+
+---
+
+<a id="html-formatting"></a>
+
+### HTML formatting
+
+*Source: <https://developers.google.com/style/html-formatting>*
+
+<a id="html-formatting--formatting"></a>
+
+Follow Google's [HTML/CSS
+Style Guide](https://google.github.io/styleguide/htmlcssguide.html). Exception: don't leave out optional elements.
+
+In particular, following are some basic guidelines from that style guide,
+which generally apply to other documentation source files, too (such as YAML and Markdown):
+
+- **Don't use tabs** to indent text; use spaces only. Different text
+  editors interpret tabs differently, and some Markdown features expect spaces
+  and not tabs.
+- **Indent by two spaces** per indentation level.
+- **Use all-lowercase** for elements and attributes.
+- **Don't leave trailing spaces** at the end of a line (except as
+  needed for Markdown).
+
+<a id="html-formatting--line-length"></a>
+
+#### Line length
+
+Break lines at 80 characters except in the following cases:
+
+- Information in a `meta` element at the beginning of a file must be on a single line,
+  so those lines can be as long as needed.
+- If a URL in a link has a line break, the link won't work.
+  If a URL is longer than 80 characters (quite common), you're stuck with it. In that case,
+  put the URL on its own line with the `href` attribute to make it
+  easier to review the text before and after, as the following example shows:
+
+```
+You can find more information in
+<a href="https://example.com/long-url/johan-gambolputty-de-von-ausfern-…-von-hautkopf-of-ulm.html"
+>his biography.</a>
+```
+
+Break code snippets (in `<pre>` blocks) at 80 characters:
+
+- Older files might use different line lengths. If you're making small changes to a file that
+  has a consistent line length other than 80 characters, then make your changes conform to that
+  file's line length rather than reformatting the whole file.
+- When adding line breaks, make sure that you don't change the meaning of the code! If you're
+  not familiar with the programming language, ask for help from someone who is. But sometimes you
+  just can't avoid a long line.
+
+---
+
+<a id="markdown"></a>
+
+### Markdown versus HTML
+
+*Source: <https://developers.google.com/style/markdown>*
+
+<a id="markdown--html-markdown"></a>
+
+Use either HTML or Markdown. Some of this style guide assumes that you're using HTML. If you're
+using Markdown, details like what HTML elements to use in various contexts might be
+irrelevant to you.
+
+Markdown is easier to write than HTML, and it's easier for most humans to
+read Markdown source than HTML source. However, HTML is more expressive
+(particularly regarding [semantic tagging](#semantic-tagging))
+and can achieve some specific effects that might be difficult or impossible in
+Markdown. For example, you might have to switch to using the HTML `code` element
+for special characters in code such as nonbreaking spaces.
+
+In the end, which one to use is primarily a matter of personal preference;
+however, if your team or your document template already uses one or the other,
+it may be best to use whatever they use.
+
+---
+
+## Names and naming
+
+<a id="examples"></a>
+
+### Example domains and names
+
+*Source: <https://developers.google.com/style/examples>*
+
+Don't use real domain names, email addresses, or people's names in your examples. Don't reveal
+personally identifiable information (PII), such as domain names, email addresses,
+phone numbers, people's names, project names, or credit card numbers. You can
+provide imaginary (fictitious) examples or use
+[placeholders](#placeholders), like
+`USER_ID` or `EMAIL_ADDRESS`.
+
+<a id="examples--domains"></a>
+
+<a id="examples--example-domain-names"></a>
+
+#### Example domain names
+
+When you need a generic domain name in an example, use example.com,
+example.org, or example.net. These domains are reserved by the
+[Internet Assigned Numbers Authority](https://www.iana.org/domains/reserved)
+for use in documentation.
+
+Alternatively, you can use any of the following domain names, which Google
+owns specifically for use in documentation:
+
+- altostrat.com
+- examplepetstore.com
+- example-pet-store.com
+- myownpersonaldomain.com
+- my-own-personal-domain.com
+- cymbalgroup.com
+
+If you need an example domain name for an internationalized domain name, use one of the
+[IDN Test TLDs](https://en.wikipedia.org/wiki/IDN_Test_TLDs) and copy from the
+"URL of the test site" column.
+
+Recommended: Hostnames that include non-ASCII characters
+are encoded using Punycode. For example, `http://مثال.إختبار` is encoded as
+`xn--kgbechtv`.
+
+<a id="examples--addresses"></a>
+
+<a id="examples--example-email-addresses"></a>
+
+#### Example email addresses
+
+If you need a generic email address, use one of the domains listed
+in
+[Example domain names](#examples--example-domain-names)
+and one of the names listed in [Example person names](#examples--example-person-names)—for
+example, dana@example.com. It's OK to use generic addresses like support@example.net. Don't use
+person names, product names, or made-up names in email addresses.
+
+<a id="examples--names"></a>
+
+<a id="examples--example-person-names"></a>
+
+#### Example person names
+
+When you need to include example given names in your documentation,
+draw from the following list:
+
+- Alex
+- Amal
+- Ariel
+- Bola
+- Charlie
+- Cruz
+- Dana
+- Dani
+- Hao
+- Ira
+- Izumi
+- Jie
+- Kai
+- Kalani
+- Kim
+- Kiran
+- Lee
+- Lucian
+- Luka
+- Mahan
+- Noam
+- Nur
+- Quinn
+- Raha
+- Rosario
+- Sasha
+- Tal
+- Taylor
+- Tristan
+- Yuri
+
+<a id="examples--example-person-surnames"></a>
+
+##### Example person surnames
+
+When you need to include example surnames in your documentation, use an initial
+after the given first name—for example, Quinn N. or Dana A.
+
+<a id="examples--further-notes-about-example-people"></a>
+
+##### Further notes about example people
+
+When you are writing about people, even fictitious or hypothetical people, it's important to
+remember that your work will be read by real people whom we want to feel respected, valued, and
+welcomed.
+
+Your audience includes different kinds of people, including people with different jobs,
+cultural contexts, and backgrounds, so strive to include a variety of people in your examples
+as well.
+
+Use the [gender-neutral singular pronouns](#pronouns--gender-neutral-pronouns)
+*they*, *their*, and *theirs*
+whenever possible, and avoid specifying gender unless it is integral to the information you
+are communicating. Avoid examples that depend on a gender binary. However, if you do write an
+example that requires specifying gender, consider that some of the names on this list may imply
+a particular gender in a given language or culture, and check to ensure that any names you have
+chosen do not carry a conflicting gender connotation.
+
+Be mindful of assumptions and stereotypes that might be reinforced through hypothetical
+examples, such as:
+
+- Job roles and levels, such as executive, that might be disproportionately assigned
+  particular gendered personas.
+- Job roles, such as developer or engineer, that might be disproportionately assigned
+  particular ethnic personas.
+
+We recommend using names from the preceding list in most documentation. Some security
+documentation uses the
+[Alice and Bob](https://wikipedia.org/wiki/Alice_and_Bob#Cast_of_characters)
+cast of characters. Don't use the Alice and Bob characters unless you're writing documentation that
+refers to a technical specification that uses those characters. If you use the Alice and Bob
+characters in a document, use only names from that cast of characters.
+
+For further guidance, see the section of this guide on
+[writing inclusive documentation](#inclusive-documentation).
+
+<a id="examples--company"></a>
+
+<a id="examples--example-company-names"></a>
+
+#### Example company names
+
+When you need a company name in an example, use Example Organization. If you need to
+differentiate between two different fictional companies, you can add a description to the company
+names. For example, you can use Enterprise Example Organization and Startup Example
+Organization.
+
+<a id="examples--phone-numbers"></a>
+
+<a id="examples--example-phone-numbers"></a>
+
+#### Example phone numbers
+
+Most phone numbers in our documentation are examples. To show an example phone number, use a US
+number in the range 800‑555‑0100 through 800‑555‑0199. That range is
+reserved for use in examples and in fiction.
+
+Never use a real phone number in examples.
+
+For information about formatting, see
+[Format phone numbers in HTML or Markdown](#phone-numbers--format-phone-numbers).
+
+<a id="examples--ip-addresses"></a>
+
+<a id="examples--example-ip-addresses"></a>
+
+#### Example IP addresses
+
+When you need an IPv4 address in an example, such as in a log, use one of the
+[RFC 5737](https://tools.ietf.org/html/rfc5737) addresses that are
+reserved for use in documentation:
+
+- `192.0.2.0` through `192.0.2.255`
+- `198.51.100.0` through `198.51.100.255`
+- `203.0.113.0` through `203.0.113.255`
+
+For IPv4 address ranges, use the following examples:
+
+- `192.0.2.0/24`
+- `198.51.100.0/24`
+- `203.0.113.0/24`
+
+When you need an IPv6 address, use values from the
+[RFC 3849](https://tools.ietf.org/html/rfc3849) range. Example IPv6 addresses include
+the following:
+
+- `2001:db8::`
+- `2001:db8:ffff:ffff:ffff:ffff:ffff:ffff`
+- `2001:db8:1:1:1:1:1:1`
+- `2001:db8:2:2:2:2:2:2`
+- `2001:db8:3:3:3:3:3:3`
+- `2001:db8:4:4:4:4:4:4`
+
+For IPv6 address ranges, use the following example:
+
+- `2001:db8::/32`
+
+<a id="examples--street-addresses"></a>
+
+<a id="examples--example-street-addresses"></a>
+
+#### Example street addresses
+
+Avoid using real street addresses in examples. Instead, use one of the following fictional
+street addresses:
+
+- 1800 Amphibious Blvd.  
+  Mountain View, CA 94045
+- Avenida da Pastelaria, 1903  
+  Lisbon, 1229-076
+- 8 Rue du Nom Fictif  
+  341 Paris
+
+<a id="examples--project-names"></a>
+
+<a id="examples--example-project-names"></a>
+
+#### Example project names
+
+When you need an example project name, create a name that's meaningful or descriptive.
+
+Ensure that the name is applicable to the reader's environment. Don't use unclear components like
+`foo`, `bar`, and `baz` in names.
+
+When necessary, use an appended numbering scheme. For example, `staging`,
+`frontend-development`, `backend-development`, `production-1`,
+`production-2`.
+
+<a id="examples--service-account-ids"></a>
+
+<a id="examples--example-service-account-ids"></a>
+
+#### Example service account IDs
+
+When you need a unique ID for a service account in an example, use the numeric ID
+`123456789012345678901`.
+
+Recommended: The allow policy shows the
+identifier `deleted:serviceAccount:my-service-account@my-project.iam.gserviceaccount.com?uid=123456789012345678901`.
+
+<a id="examples--data"></a>
+
+---
+
+<a id="filenames"></a>
+
+### Filenames
+
+*Source: <https://developers.google.com/style/filenames>*
+
+<a id="filenames--naming-guidelines"></a>
+
+#### Guidelines for names
+
+Make file and directory names lowercase, with the occasional exception for consistency, to make file searches easier and search results more useful. For example, because most Unix-style operating systems are case sensitive, they can't find a file named `Impersonate-Service-Accounts.html` if you search for `impersonate-service-accounts.html`. Linux and macOS interpret these as two distinct files.
+
+Use hyphens, not underscores, to separate words—for example,
+`query-data.html`. Search engines interpret hyphens in file and directory names as spaces between words. Underscores are generally not recognized, meaning that their presence can negatively affect SEO.
+
+Use only standard ASCII
+alphanumeric characters in file and directory names.
+
+Don't use generic page names such as `document1.html`.
+
+<a id="filenames--exceptions-for-consistency"></a>
+
+##### Exceptions for consistency
+
+If you're adding to a directory where everything else already uses
+underscores, and it's not feasible to change everything to hyphens, it's okay to
+use underscores to stay consistent.
+
+For example, if the directory already has `lesson_1.jd`,
+`lesson_2.jd`, and `lesson_3.jd`, it's okay to add your
+new file as `lesson_4.jd` instead of `lesson-4.jd`.
+However, in all other situations, use hyphens.
+
+Recommended: `avoiding-cliches.jd`
+
+Sometimes OK: `avoiding_cliches.jd`
+
+Not recommended: `avoidingcliches.jd`
+
+Not recommended: `avoidingCliches.jd`
+
+Not recommended: `avoiding-clichés.jd`
+
+<a id="filenames--other-exceptions"></a>
+
+##### Other exceptions
+
+It's okay to have some inconsistency in filenames if it can't otherwise be
+avoided. For example, sometimes tools that generate reference documentation
+produce filenames based on different style requirements or based on the design
+and naming conventions of the product or API itself. In those cases, it's okay
+to make exceptions for those files.
+
+<a id="filenames--referring-files"></a>
+
+#### Refer to files
+
+The following sections discuss how to reference files.
+
+<a id="filenames--referring-filenames"></a>
+
+##### Refer to filenames
+
+When referring to a specific file, do the following:
+
+- Use [code font](#code-in-text).
+- Include the word *file* after the filename. For more information, see
+  [Grammatical treatment of code elements](#code-in-text--grammatical-treatment-of-code-elements).
+- Use the exact spelling of the filename even if it doesn't follow
+  [naming guidelines](#filenames--naming-guidelines).
+- If a sample of the file is included on the page, follow the
+  [code sample](#code-samples)
+  guidelines and precede a code sample with an introductory sentence or paragraph that includes the
+  filename.
+
+Recommended: In the following
+`build.sh` file, modify the default values for all parameters:
+
+<a id="filenames--file-type-interactions"></a>
+
+##### Refer to file interactions
+
+When interacting with files and file types, don't use the file types as a verb.
+
+Recommended: Extract a zip file.
+
+Not recommended: Unzip a zip file.
+
+<a id="filenames--file-type-names"></a>
+
+##### Refer to file types
+
+When you're discussing a file type, use the formal name of the type, not the filename extension.
+(The file type name is often in all caps because many file type names are acronyms
+or initialisms.) Do not use the filename extension to refer generically to the
+file type.
+
+Recommended: a PNG file
+
+Not recommended: a `.png`
+file
+
+Recommended: a Bash file
+
+Not recommended: an `.sh`
+file
+
+The following table lists some examples of filename extensions and the
+corresponding file type names to use.
+
+| Extension | File type name |
+| --- | --- |
+| `.adoc` | AsciiDoc file |
+| `.csv` | CSV file |
+| `.exe` | executable file |
+| `.gif` | GIF file |
+| `.img` | disk image file |
+| `.ipynb` | IPYNB file |
+| `.jar` | JAR file |
+| `.jpg`, `.jpeg` | JPEG file |
+| `.json` | JSON file |
+| `.md` | Markdown file |
+| `.pdf` | PDF file |
+| `.png` | PNG file |
+| `.ps` | PowerShell file |
+| `.py` | Python file |
+| `.sh` | Bash file |
+| `.sql` | SQL file |
+| `.svg` | SVG file |
+| `.tar` | tar file |
+| `.tf` | Terraform file |
+| `.tiff` | TIFF file |
+| `.txt` | text file |
+| `.wasm` | Wasm file |
+| `.yaml` | YAML file |
+| `.zip` | zip file |
+
+---
+
+<a id="trademarks"></a>
+
+### Trademarks
+
+*Source: <https://developers.google.com/style/trademarks>*
+
+Follow any usage guidelines that trademark owners provide.
+<a id="trademarks--labeling"></a>
+
+<a id="trademarks--label-trademarked-terms"></a>
+
+#### Label trademarked terms
+
+For trademark marking or attribution in documentation, follow any usage
+guidelines provided by the owners of the respective marks.
+
+For more about Google trademarks in particular, see [About our trademarks and
+how to use them](https://www.google.com/permissions/trademark/).
+
+<a id="trademarks--adjectives"></a>
+
+<a id="trademarks--use-trademarks-only-as-modifiers"></a>
+
+#### Use trademarks only as modifiers
+
+When you use a trademarked term, always use it to modify a noun, not as a noun
+by itself. Don't use a trademark as a verb.
+
+Never form a possessive or a plural from a trademark or change it in any way. For more
+information, see [Possessives](#possessives).
+
+Recommended: Another option is to use a Chromebook notebook computer.
+
+Not recommended: Another option is to use a Chromebook.
+
+Not recommended: Chromebook's features rely on an internet connection.
+
+Not recommended: For information about Chromebook computers, google "notebook computers"
+
+For more information about using Google trademarks, see [Rules for proper
+usage](https://www.google.com/permissions/trademark/rules.html).
diff --git a/pyproject.toml b/pyproject.toml
index 88f39b8..48add70 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -8,6 +8,12 @@ dependencies = ["marker-pdf>=1.10.2", "psutil>=7.2.2"]
 
 [dependency-groups]
 dev = ["ruff>=0.8"]
+web2md = [
+  "httpx>=0.28.1",
+  "beautifulsoup4>=4.15.0",
+  "lxml>=6.1.1",
+  "markdownify>=1.2.3",
+]
 
 # No [build-system] / src layout yet — there is no first-party Python package in
 # this repo. Adopt src/md2okf/ + tests/ + pytest when real Python code lands
diff --git a/uv.lock b/uv.lock
index 6ef673d..7de2d79 100644
--- a/uv.lock
+++ b/uv.lock
@@ -649,6 +649,86 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/7b/91/984aca2ec129e2757d1e4e3c81c3fcda9d0f85b74670a094cc443d9ee949/joblib-1.5.3-py3-none-any.whl", hash = "sha256:5fc3c5039fc5ca8c0276333a188bbd59d6b7ab37fe6632daa76bc7f9ec18e713", size = 309071, upload-time = "2025-12-15T08:41:44.973Z" },
 ]
 
+[[package]]
+name = "lxml"
+version = "6.1.1"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/05/3b/aab6728cae887456f409b4d75e8a01856e4f04bd510de38052a47768b680/lxml-6.1.1.tar.gz", hash = "sha256:ba96ae44888e0185281e937633a743ea90d5a196c6000f82565ebb0580012d40", size = 4197430, upload-time = "2026-05-18T19:19:06.424Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/6a/6e/c4add832b6fc1e887125b96f880d7b9b70aae5248718e046b1704bcac4b9/lxml-6.1.1-cp312-cp312-macosx_10_13_universal2.whl", hash = "sha256:104c09bda8d2a562824c0e319d0768ce26a779b7601e0931d33b09b53c392ef7", size = 8570821, upload-time = "2026-05-18T19:17:42.068Z" },
+    { url = "https://files.pythonhosted.org/packages/22/00/ff3009c88e65de8011630acf8ab5a09cb2becd2aaf47fba2f3449f6224e9/lxml-6.1.1-cp312-cp312-macosx_10_13_x86_64.whl", hash = "sha256:25c6997a9a534e016695a0ba06b2f07945de682731ff01065b6d5a4474179da1", size = 4624252, upload-time = "2026-05-18T19:17:47.897Z" },
+    { url = "https://files.pythonhosted.org/packages/42/95/bb63f0fd62e554fe078e1fb3c8fe9083c14ddc7ad7fa178d10e57e071ac7/lxml-6.1.1-cp312-cp312-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:c921ba5c51e4e9f63b8b00267d06566e1f63407408a0496da2d1d0bfc819c7fc", size = 4930746, upload-time = "2026-05-18T19:18:29.637Z" },
+    { url = "https://files.pythonhosted.org/packages/eb/99/0013e8d9b5960f4f041cf0b73e2f80c23eb5205b1f7bfb20203243651359/lxml-6.1.1-cp312-cp312-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:54a7f95e4de5fb94e2f9f4b9055c6ba33bf3d628fd77a1d647c5923caa2cdcdc", size = 5093723, upload-time = "2026-05-18T19:18:34.168Z" },
+    { url = "https://files.pythonhosted.org/packages/29/91/317b332636bfc7bddcff828d41b3307f50043f4b237e40849c333d80fa1a/lxml-6.1.1-cp312-cp312-manylinux_2_26_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:96f2ec43df44b1f76249ee0a615334f9b5b060e1c8bd90e706dad2d14d02f383", size = 5005557, upload-time = "2026-05-18T19:18:39.798Z" },
+    { url = "https://files.pythonhosted.org/packages/42/2f/cc9bf06afe70f9c9093ae60855d9759da9db601ec4080f7473319666ffd7/lxml-6.1.1-cp312-cp312-manylinux_2_26_ppc64le.manylinux_2_28_ppc64le.whl", hash = "sha256:70ef8a7e102a1508f8121aae5b0867abd663f72c14f0a9c937e6554cb4587b7b", size = 5631036, upload-time = "2026-05-18T19:18:44.858Z" },
+    { url = "https://files.pythonhosted.org/packages/08/f6/af32e23e563971ffb0fb86be52bc5be5c2c118858ffc119bf6a9039b173d/lxml-6.1.1-cp312-cp312-manylinux_2_26_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:ebe6af670449830d6d9b752c256a983291c766a1365ba5d5460048f9e33a7818", size = 5240367, upload-time = "2026-05-18T19:18:49.217Z" },
+    { url = "https://files.pythonhosted.org/packages/78/83/8555d40948b09ce86f1bd0c68a7ac31d07b1929f92cc1b074006c97ef2d2/lxml-6.1.1-cp312-cp312-manylinux_2_28_i686.whl", hash = "sha256:27acc820660aaffa4f7c087f29120e12980f7779d56d8492d263170111284740", size = 5350171, upload-time = "2026-05-18T19:18:52.779Z" },
+    { url = "https://files.pythonhosted.org/packages/63/75/5d92da93729b7bad783689e6496049fa40927b45bec7bf183c981de3ca70/lxml-6.1.1-cp312-cp312-manylinux_2_31_armv7l.whl", hash = "sha256:1db753c9115ec7100d073b744d17e25e88a8f90f5c39b2f5dd878149af59671f", size = 4694874, upload-time = "2026-05-18T19:18:55.139Z" },
+    { url = "https://files.pythonhosted.org/packages/c5/b5/3aad415a9a25b822e783f15deeb4dffccf5113030f1afa2222dd929313d9/lxml-6.1.1-cp312-cp312-manylinux_2_38_riscv64.manylinux_2_39_riscv64.whl", hash = "sha256:c4f469aebd783bb741c2ecb2a681008fd26bfe5c16a9a72ed5467f834e810df2", size = 5244492, upload-time = "2026-05-18T19:19:01.28Z" },
+    { url = "https://files.pythonhosted.org/packages/f1/a1/5fcf7eb9904b80086aa47dcf0027de07b1bb990afad2e6823144c368ae04/lxml-6.1.1-cp312-cp312-musllinux_1_2_aarch64.whl", hash = "sha256:766b010012d59470072c1816b5b6c69f1d243e5db36ea5968e94accf430a4635", size = 5048232, upload-time = "2026-05-18T19:18:12.67Z" },
+    { url = "https://files.pythonhosted.org/packages/77/74/1f601b63c7a69fcdf10fa9b148c81da8442204194f6c55509cc485c786b9/lxml-6.1.1-cp312-cp312-musllinux_1_2_armv7l.whl", hash = "sha256:b8d812c6011c08b8111a15e54dd990b8923692d80adf35488bee34026c35accf", size = 4777023, upload-time = "2026-05-18T19:18:15.928Z" },
+    { url = "https://files.pythonhosted.org/packages/a2/b9/7a78f51aec95b1bf780d78e12705a9f6533284f8693dc5c0e6724fa53d3f/lxml-6.1.1-cp312-cp312-musllinux_1_2_ppc64le.whl", hash = "sha256:fe0306bd29505a9177aac19f1877174b0e7422c222a59f70b2cd41633448c3dc", size = 5645773, upload-time = "2026-05-18T19:18:23.223Z" },
+    { url = "https://files.pythonhosted.org/packages/a5/6e/98a7b7ad54e4e74fa1f20fff776913980619d0ebe5558232d7da6580bdd8/lxml-6.1.1-cp312-cp312-musllinux_1_2_riscv64.whl", hash = "sha256:5ba186ad207446c65d3bb3d3e0412b032b1d9f595e59861e2354798c5703d955", size = 5233088, upload-time = "2026-05-18T19:18:31.433Z" },
+    { url = "https://files.pythonhosted.org/packages/65/d1/bc0ed2427bf609f2ee10da303a6a226f9c8bce94f945dc29a32ce55de6e4/lxml-6.1.1-cp312-cp312-musllinux_1_2_x86_64.whl", hash = "sha256:aa366a1e55b8ebfe8ca8ddc3cfe75c8ebade181aeb0f661d0cb05986b647f72a", size = 5260995, upload-time = "2026-05-18T19:18:37.091Z" },
+    { url = "https://files.pythonhosted.org/packages/69/8b/6772e1a4b513fc50a8d931f19edde0e13ae6918510a1e13ff67864f3e5ed/lxml-6.1.1-cp312-cp312-win32.whl", hash = "sha256:126c93f7f56f0eda92f6d8c619edc463a4f23d9252f1c9d0405a76f25fa9f11a", size = 3596382, upload-time = "2026-05-18T19:17:18.37Z" },
+    { url = "https://files.pythonhosted.org/packages/1b/89/45198e9624762af2dfd2cb8782598477ceb29f6e59caab560388ae1f4ec1/lxml-6.1.1-cp312-cp312-win_amd64.whl", hash = "sha256:26e6eda8d38c1fcab1090dd196ee87cbd13788e531937610e2589085de074e77", size = 3997255, upload-time = "2026-05-18T19:17:56.781Z" },
+    { url = "https://files.pythonhosted.org/packages/90/a9/7a54b6834088d9ae528a7b780584ba6a39a9457b0ac330479f20ffbc9449/lxml-6.1.1-cp312-cp312-win_arm64.whl", hash = "sha256:6540377fbd53fe1b629172288c464fb18db11ce1fa7dc15891da10aa9dcc3e7f", size = 3659610, upload-time = "2026-05-19T19:22:50.843Z" },
+    { url = "https://files.pythonhosted.org/packages/a5/eb/7e6f37c5584ccbb2ff267f56fd0339016938c1c8684cfefab9b33ffc2f36/lxml-6.1.1-cp313-cp313-macosx_10_13_universal2.whl", hash = "sha256:68a9198d0fc122d14bb76837de9aa80cf84caed990b5b237f532ed87d3706736", size = 8559780, upload-time = "2026-05-18T19:17:57.661Z" },
+    { url = "https://files.pythonhosted.org/packages/a1/36/587c2521cf23a2cd6c9c22108aa7528f683a1f195ed7ccd23a4b1786ad36/lxml-6.1.1-cp313-cp313-macosx_10_13_x86_64.whl", hash = "sha256:7d47866cb32fb503450b6edc9df355d10dc49836af2e89901bd6ac6b0896d9d9", size = 4618006, upload-time = "2026-05-18T19:18:04.452Z" },
+    { url = "https://files.pythonhosted.org/packages/6e/ca/ab7bfe2bf4c972af5e7878262845ead3a24a929a9b04bc11c7c1ece6c82a/lxml-6.1.1-cp313-cp313-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:eb7c9811bfaa8b1ed5ed319f5d370dfbcaa59d52ea64be2a5a85e18195930354", size = 4924139, upload-time = "2026-05-18T19:19:04.873Z" },
+    { url = "https://files.pythonhosted.org/packages/6b/55/a0c72851dfee5ecc689f949723a73dea457758912542cb955b108eaf0d8f/lxml-6.1.1-cp313-cp313-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:762ff394d5bd56da0cf034a23dcce4e13923f15321a2adfa2ac00201dc6d3fca", size = 5082329, upload-time = "2026-05-18T19:19:09.728Z" },
+    { url = "https://files.pythonhosted.org/packages/f0/b6/0608f7d61a3b96cc67e5648a3d906e31a5082093e10e7be65b3886289938/lxml-6.1.1-cp313-cp313-manylinux_2_26_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:a088f287f7d8275a33c07f2cac6c50b9319309a0200a39e7e75d80c707723099", size = 4993564, upload-time = "2026-05-18T19:19:13.608Z" },
+    { url = "https://files.pythonhosted.org/packages/4c/66/ae227524b066d29d55bf0b453d93d2d793c40218657d643dcbbca13b8faf/lxml-6.1.1-cp313-cp313-manylinux_2_26_ppc64le.manylinux_2_28_ppc64le.whl", hash = "sha256:e902da4b04e6b52e5893900d4b8ab46068f75f3561f01bf1080957f9fd932ed6", size = 5613467, upload-time = "2026-05-18T19:19:16.228Z" },
+    { url = "https://files.pythonhosted.org/packages/a6/76/dbe4a00b50385e40194231dcfe5a12c059de7cf90e89c83407d2b085b719/lxml-6.1.1-cp313-cp313-manylinux_2_26_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:1d4962d4c66bf830a7e59ed6cfc17d148149898a3aefa8ec6e59763e6e3ed085", size = 5228304, upload-time = "2026-05-18T19:19:19.354Z" },
+    { url = "https://files.pythonhosted.org/packages/1c/01/00b1b8442ed2041793336868ba0b9ea4b13d7da7c085c6404c207a63bf79/lxml-6.1.1-cp313-cp313-manylinux_2_28_i686.whl", hash = "sha256:581d4c8ae690a6609e64862dd6b7c2489635c2d13907fc2b20f2bc200ff1d21e", size = 5341607, upload-time = "2026-05-18T19:19:22.297Z" },
+    { url = "https://files.pythonhosted.org/packages/63/36/1ad29931e9a4638bb707869f01d423a6c815f82152138d1a40dfcfde2b95/lxml-6.1.1-cp313-cp313-manylinux_2_31_armv7l.whl", hash = "sha256:876e1ff5930ed8bf295ec5ef9a8155e9b6b1876bbf1deed8b3a8069311875a8f", size = 4700168, upload-time = "2026-05-18T19:19:25.133Z" },
+    { url = "https://files.pythonhosted.org/packages/3c/d1/a9536cecf9be18a0dc72d32bead283a2332d1ffebd2dd3ac70ce444686e5/lxml-6.1.1-cp313-cp313-manylinux_2_38_riscv64.manylinux_2_39_riscv64.whl", hash = "sha256:9eb9b5a968f6e0f6d640092a567e14529ff8cea2e29d00da6f78a79fa49f013c", size = 5232487, upload-time = "2026-05-18T19:19:28.603Z" },
+    { url = "https://files.pythonhosted.org/packages/0e/77/b4fb1e03bf5d130e879214d3100092e386418807fb74dd0adc4b0a48f351/lxml-6.1.1-cp313-cp313-musllinux_1_2_aarch64.whl", hash = "sha256:aa49e06d94aba782c6a02eecb7e507969e7e7a41b267f1b359bb35585f295d5b", size = 5044231, upload-time = "2026-05-18T19:18:42.246Z" },
+    { url = "https://files.pythonhosted.org/packages/26/4c/d00daeeb0a5530c4028a9232aa1b93db3ef4ed2158c116ea73c79a9765b3/lxml-6.1.1-cp313-cp313-musllinux_1_2_armv7l.whl", hash = "sha256:70cdfd80589d59e43e18005dd7244e8895e93db8ab6a620b7e23df5445a4e3d2", size = 4769450, upload-time = "2026-05-18T19:18:48.013Z" },
+    { url = "https://files.pythonhosted.org/packages/ed/6a/715a3a8d156ce42f29cf014706f5410c2ff3b02267774110fc23266409fe/lxml-6.1.1-cp313-cp313-musllinux_1_2_ppc64le.whl", hash = "sha256:aad9aa39483ed8ec44d6d2e59e5b98a0d80676ef0d92f44bfc374836111f62f5", size = 5635874, upload-time = "2026-05-18T19:18:51.914Z" },
+    { url = "https://files.pythonhosted.org/packages/45/37/0544bc21dde2a88f3a17b504e6fc79c0e01d25a33c2f6079724e9e72b9c7/lxml-6.1.1-cp313-cp313-musllinux_1_2_riscv64.whl", hash = "sha256:d49514be2f28d895c38cf9d2b72d7b9a07d00314519f456c0b50b53cfcf4c785", size = 5223987, upload-time = "2026-05-18T19:18:59.715Z" },
+    { url = "https://files.pythonhosted.org/packages/4d/f8/f6a5e8185bcb28c2befae3d31f8e3df3b811cb0f47746517a81279fcafe1/lxml-6.1.1-cp313-cp313-musllinux_1_2_x86_64.whl", hash = "sha256:47402e62c52ff5988c1e8c6c63177f5708bccf48e366dea4e3dcf1e645e04947", size = 5250276, upload-time = "2026-05-18T19:19:03.834Z" },
+    { url = "https://files.pythonhosted.org/packages/c7/f2/1a2b9f1b7a49d45495369be7ef9ad05b262930f2eab3e3145706fca8083f/lxml-6.1.1-cp313-cp313-win32.whl", hash = "sha256:3483644525531e1d5762b0c44a8e18b6efba321b6dcf8a8952de10b037618bca", size = 3596903, upload-time = "2026-05-18T19:17:29.863Z" },
+    { url = "https://files.pythonhosted.org/packages/e6/99/f4ffb024f238eec2131aaa09f3278fb6129cf892741bf68e1fc1afb8c100/lxml-6.1.1-cp313-cp313-win_amd64.whl", hash = "sha256:a10bd2fd62e8ce916ececb342f348f190724a098c1faa056fdfb2a22ad5e8660", size = 3995869, upload-time = "2026-05-18T19:18:02.596Z" },
+    { url = "https://files.pythonhosted.org/packages/d1/53/70eb8c5c6037f27448f1e3c54ebede9545a801ae63f0a7254afca4fe8e45/lxml-6.1.1-cp313-cp313-win_arm64.whl", hash = "sha256:424aa57aca0897eb922aef34395bd1289b3b6f04e6bae20ea123c0c7e333cffc", size = 3658490, upload-time = "2026-05-19T19:22:53.846Z" },
+    { url = "https://files.pythonhosted.org/packages/13/e2/2e325795566de01d0d7c3bb57d3c370616b2d07b01214e84eec5d3b10963/lxml-6.1.1-cp314-cp314-macosx_10_15_universal2.whl", hash = "sha256:19b7ab10b210b0b3ad7985d9ac4eb66ab09a90b20fe6e2f7ba55d01a234345d0", size = 8577146, upload-time = "2026-05-18T19:18:17.765Z" },
+    { url = "https://files.pythonhosted.org/packages/93/cf/5630b5e4be7d2e6bee8efe83865c925221103cf0221303b104ce134b01e2/lxml-6.1.1-cp314-cp314-macosx_10_15_x86_64.whl", hash = "sha256:c08e5c694306507275f2290073350c4f32e383db15213b2c69e7ff39c1193840", size = 4623866, upload-time = "2026-05-18T19:18:30.669Z" },
+    { url = "https://files.pythonhosted.org/packages/d2/51/3904907c063451cf8d4a5c9fe0cad95fa1f4ec57f4e3884fa0731bd7a305/lxml-6.1.1-cp314-cp314-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:74a9717fd0d82effef5c2854f0d917231d5324b5a3eb7275c43ac9fa32f97a14", size = 4950022, upload-time = "2026-05-18T19:19:31.958Z" },
+    { url = "https://files.pythonhosted.org/packages/94/cd/9c7611a51c37a2830928405817cc5d56a97f64fab83cc3f628748b135749/lxml-6.1.1-cp314-cp314-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:efe0374196335f93b53269acd811b944f2e6bdc88e8894f214bd636455484909", size = 5086695, upload-time = "2026-05-18T19:19:34.764Z" },
+    { url = "https://files.pythonhosted.org/packages/da/d6/24e3b5906abb0b674ff2ae195bc3ce59708df2bcd17cf17703b2d7dd643a/lxml-6.1.1-cp314-cp314-manylinux_2_26_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:ac931cdc9442c1763b8a8f6cd62c0c938737eafc5be75eff88df55fc73bc0d00", size = 5031642, upload-time = "2026-05-18T19:19:37.771Z" },
+    { url = "https://files.pythonhosted.org/packages/2d/db/6ec54f99019838bff54785c51da07f189eb4676861c5f2730962b0d8d665/lxml-6.1.1-cp314-cp314-manylinux_2_26_ppc64le.manylinux_2_28_ppc64le.whl", hash = "sha256:aee395f5d0927f947758b4ec119fd5fc8ec71f07a1c5c52077b30b04c0fa6955", size = 5647338, upload-time = "2026-05-18T19:19:40.553Z" },
+    { url = "https://files.pythonhosted.org/packages/42/3d/ef4dcfffd22d27a61805d8ed9f7fb888495bc6aa88648fa07c1eaa5586b6/lxml-6.1.1-cp314-cp314-manylinux_2_26_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:9395002973c827b3ed67db77e6ec09f092919a587022174554096a269378fb13", size = 5239528, upload-time = "2026-05-18T19:19:43.657Z" },
+    { url = "https://files.pythonhosted.org/packages/62/bb/37fb3f0dff146bdcfa78eec47879273820b2a0bf350ec236ce14bd0b1c26/lxml-6.1.1-cp314-cp314-manylinux_2_28_i686.whl", hash = "sha256:73bc2086f141224ebddb7fc5c6a36ca58b31b94b561e1dfe8e073e3270fad1e7", size = 5350730, upload-time = "2026-05-18T19:19:46.307Z" },
+    { url = "https://files.pythonhosted.org/packages/90/42/43253f168388df4fae1f38c01df36ddb9bee39e2048167b54cdcbae85ea3/lxml-6.1.1-cp314-cp314-manylinux_2_31_armv7l.whl", hash = "sha256:3779def59032b81e44a5f70096ef6bf2082f8d901937dca354474ba09782e245", size = 4697530, upload-time = "2026-05-18T19:19:49.889Z" },
+    { url = "https://files.pythonhosted.org/packages/eb/a8/c5a8504f81bbdfc8e7094c2c850cdb4ed6777fc4d5ddd9e5ab819f3b0d54/lxml-6.1.1-cp314-cp314-manylinux_2_38_riscv64.manylinux_2_39_riscv64.whl", hash = "sha256:86c89b9d55ebf820ad7c90bc533410f0d098054f293351f10603c0c46ff598f5", size = 5250670, upload-time = "2026-05-18T19:19:53.199Z" },
+    { url = "https://files.pythonhosted.org/packages/77/b7/c7e76ab18744d75e21f320ebf9ff9d1ceae2b54dd431ea5a64caf26c9672/lxml-6.1.1-cp314-cp314-musllinux_1_2_aarch64.whl", hash = "sha256:19607c6bbff2a44cf3fe8250abccd20942d3462473e0a721d01d379ed017e462", size = 5084485, upload-time = "2026-05-18T19:19:08.422Z" },
+    { url = "https://files.pythonhosted.org/packages/31/31/b35c53f8ef7b7c31cacd23d3638652fff7bcd1deb6eedb709ab43b685908/lxml-6.1.1-cp314-cp314-musllinux_1_2_armv7l.whl", hash = "sha256:c6ed5141a5c7507cf3ee76bd363b0d6f801e3321adc35b5d825a23115faa5465", size = 4737635, upload-time = "2026-05-18T19:19:12.321Z" },
+    { url = "https://files.pythonhosted.org/packages/d9/06/31f23c813a7fe8e0cb1b175e915b08c9bf4e86d225b210feadbdbe519667/lxml-6.1.1-cp314-cp314-musllinux_1_2_ppc64le.whl", hash = "sha256:62aeb7e85b5d60320b9d77eef2e773994e2c0ce10121b277e0a19804e1654a5a", size = 5670681, upload-time = "2026-05-18T19:19:15.001Z" },
+    { url = "https://files.pythonhosted.org/packages/1a/bc/ce619bccc89b1fd9ad8a8e1330ee3f3beff9f2ff95b712d7bbcdd6e22fc3/lxml-6.1.1-cp314-cp314-musllinux_1_2_riscv64.whl", hash = "sha256:b1b963fd8f5caa68e99dfae060d54de1fe9cba899b8718b44a00cdca53c3e590", size = 5238229, upload-time = "2026-05-18T19:19:18.131Z" },
+    { url = "https://files.pythonhosted.org/packages/2f/5d/b329acbbedc0b619ebc2be6cf7ee9ed07e80892c88d4dfd612c33805789a/lxml-6.1.1-cp314-cp314-musllinux_1_2_x86_64.whl", hash = "sha256:63876be28efefa04a1df615b46770e82042cce445cfdce55160522f57b231ccb", size = 5264191, upload-time = "2026-05-18T19:19:21.118Z" },
+    { url = "https://files.pythonhosted.org/packages/d6/85/be36fb1425b30db3c3f9df75fe86343ebffb79e6320bd7f588e25bfeac39/lxml-6.1.1-cp314-cp314-win32.whl", hash = "sha256:7f7a92e8583f06b1fd49d01158143b8461cfcd135dcb10ec807270a3051bd603", size = 3657202, upload-time = "2026-05-18T19:17:39.509Z" },
+    { url = "https://files.pythonhosted.org/packages/b8/ce/3cf9a827342269f54d405a6202397de63f07c69cbd6ce7d183a3f0cba1e9/lxml-6.1.1-cp314-cp314-win_amd64.whl", hash = "sha256:b2d444f2e66624d68e9c6b211e28a76e22fff5fcabcfff4deac18b529b7d4137", size = 4064497, upload-time = "2026-05-18T19:18:14.662Z" },
+    { url = "https://files.pythonhosted.org/packages/d9/3e/1a957bde8f0760039e627f94699f82caa782c9d838d86c3d28245ee67212/lxml-6.1.1-cp314-cp314-win_arm64.whl", hash = "sha256:3fd9728a2735fda14f4e8235830c86b539e9661e849665bf926d3f867943b4bf", size = 3741991, upload-time = "2026-05-19T19:22:59.111Z" },
+    { url = "https://files.pythonhosted.org/packages/78/b2/00ed55b3a2efa4658fb795c38d1090ec9b3e8a6c3683d4441fa517f09c3b/lxml-6.1.1-cp314-cp314t-macosx_10_15_universal2.whl", hash = "sha256:787b2496d0dbe8cd180984e8d29e3a6f76e7ea34db781cb3bd55e4ba1ef8b4ee", size = 8827545, upload-time = "2026-05-18T19:18:41.193Z" },
+    { url = "https://files.pythonhosted.org/packages/c0/73/74573db19baa618d5f266f2407898b087ff6927115b00b71e5fc1b700847/lxml-6.1.1-cp314-cp314t-macosx_10_15_x86_64.whl", hash = "sha256:2c8daa471358dc2d6fcf02165e80ec68f77871a286df95bc5cc3816153b0fd2c", size = 4735736, upload-time = "2026-05-18T19:18:46.761Z" },
+    { url = "https://files.pythonhosted.org/packages/16/02/6f7061f4f95f51e545d48e87647c54791d204a4e881be4156e7a26ba5338/lxml-6.1.1-cp314-cp314t-manylinux2014_aarch64.manylinux_2_17_aarch64.whl", hash = "sha256:acd7d70b64c0aae0c7922cca83d288a16f5f6da523637697872253415269baef", size = 4970291, upload-time = "2026-05-18T19:19:56.215Z" },
+    { url = "https://files.pythonhosted.org/packages/b0/02/55fc057d8283427dea7d6edb102e7a840239c77a64a983d92f62a304c0e9/lxml-6.1.1-cp314-cp314t-manylinux2014_x86_64.manylinux_2_17_x86_64.whl", hash = "sha256:4f0dd2f01f9f8a89f565d000e03abcf0a13d692a346c8d22f628d49af098777a", size = 5102822, upload-time = "2026-05-18T19:19:59.223Z" },
+    { url = "https://files.pythonhosted.org/packages/e4/48/8e1cf78d89d66850121d9255a2a24414c98f775da93b90cf976956c24b14/lxml-6.1.1-cp314-cp314t-manylinux_2_26_aarch64.manylinux_2_28_aarch64.whl", hash = "sha256:0b7e8a14c8634bf6f7a568634cb395305a6d964aeb5b7ee32248094bed3a7e2c", size = 5027923, upload-time = "2026-05-18T19:20:01.549Z" },
+    { url = "https://files.pythonhosted.org/packages/ed/00/0632a0647612c8af24d26997b3b961397daa9d5b2581444805933629a4cb/lxml-6.1.1-cp314-cp314t-manylinux_2_26_ppc64le.manylinux_2_28_ppc64le.whl", hash = "sha256:86281fbdd6a8162756f8d603f37e3435bfa38043adb79c6dc6a2dfee065e7525", size = 5595843, upload-time = "2026-05-18T19:20:03.93Z" },
+    { url = "https://files.pythonhosted.org/packages/bc/86/ab008a7dc360711b66858d61c80a5979a70a09f2aa2b05d9698df80b803d/lxml-6.1.1-cp314-cp314t-manylinux_2_26_x86_64.manylinux_2_28_x86_64.whl", hash = "sha256:c5d7152ec39ca7c402d8fb9bad86140a15b9503bd0c54484e3f1bbe3dd37ceca", size = 5224515, upload-time = "2026-05-18T19:20:06.381Z" },
+    { url = "https://files.pythonhosted.org/packages/75/c6/2702ff375e728e34f56d9a45339a9cf7e4427e917f542225242d63a05afa/lxml-6.1.1-cp314-cp314t-manylinux_2_28_i686.whl", hash = "sha256:88d8cb75b9d82858497a5393e3c63cfbf03035225e4b35a49ed7ccb151e4dc0e", size = 5312511, upload-time = "2026-05-18T19:20:09.308Z" },
+    { url = "https://files.pythonhosted.org/packages/b7/57/a5807c98f87a86f10ef9ffab35516df7c0f0c4b6d5d33e9f608ab9c04a31/lxml-6.1.1-cp314-cp314t-manylinux_2_31_armv7l.whl", hash = "sha256:f64ec5397ea6a41fc1b4af0380d79b44a755b5531dcaccd9940fb260dca93038", size = 4639206, upload-time = "2026-05-18T19:20:11.704Z" },
+    { url = "https://files.pythonhosted.org/packages/1f/e1/8a0a2c35734812395f4da4eaf33748a7e5705bfb2a58b128da764339d5ec/lxml-6.1.1-cp314-cp314t-manylinux_2_38_riscv64.manylinux_2_39_riscv64.whl", hash = "sha256:d34bbf07dbc7ca5970671b1512e928991fb5e9d95365636c9b2d8b4f53af405e", size = 5232404, upload-time = "2026-05-18T19:20:14.064Z" },
+    { url = "https://files.pythonhosted.org/packages/c2/e2/0e6a4dd5ad84d01d99aa7bae7cfefd4a760a0e0f8176818241de17d9b6c0/lxml-6.1.1-cp314-cp314t-musllinux_1_2_aarch64.whl", hash = "sha256:17e0e18d4ad8adbd0399291bc44845b69d9dd68439a3cdebdf35ff902ec05072", size = 5083769, upload-time = "2026-05-18T19:19:23.758Z" },
+    { url = "https://files.pythonhosted.org/packages/a0/7e/161f33d463f6ffc1c7679104b65086dea120080d49dde4d238f015aaee2f/lxml-6.1.1-cp314-cp314t-musllinux_1_2_armv7l.whl", hash = "sha256:3ab541146f1f6968c462d6c2ac495148e8cdba2f8347700b2141b6ec5a75bf52", size = 4758936, upload-time = "2026-05-18T19:19:27.256Z" },
+    { url = "https://files.pythonhosted.org/packages/f1/fb/2369825e3f6ca99305bf9f7b7085fda91c8b0922a89e54d900974aa3ef85/lxml-6.1.1-cp314-cp314t-musllinux_1_2_ppc64le.whl", hash = "sha256:2a0217714657e023ef4293500f65aa20fce6164c8fd6b08fa5bd4a859fb14b9b", size = 5620296, upload-time = "2026-05-18T19:19:29.993Z" },
+    { url = "https://files.pythonhosted.org/packages/30/90/d61e383146f74c5ab683947ea14dc7b82778838ab9b95ea73a23b60d0191/lxml-6.1.1-cp314-cp314t-musllinux_1_2_riscv64.whl", hash = "sha256:05a82eb6e1530a64f26225b55cbd178113bd0b5af1c2b625f25e5296742c26d2", size = 5228598, upload-time = "2026-05-18T19:19:33.523Z" },
+    { url = "https://files.pythonhosted.org/packages/76/2d/2dafd8149e94b05bb070690efd5bb2680720681e03ff03fc57d2b70a1105/lxml-6.1.1-cp314-cp314t-musllinux_1_2_x86_64.whl", hash = "sha256:9e36f163528fc50cbef305f02a5fd66d404edf7049cdaff211dbc2cba5a7013e", size = 5247845, upload-time = "2026-05-18T19:19:36.649Z" },
+    { url = "https://files.pythonhosted.org/packages/ce/68/b30e913340c380ddac9580c6e6230991fc37240ec4f64704833e4f3e2769/lxml-6.1.1-cp314-cp314t-win32.whl", hash = "sha256:649dda677cf3bd6ac9ae14007ba0c824ded8ce5808b53fc7431d9140399118c1", size = 3897345, upload-time = "2026-05-18T19:17:33.562Z" },
+    { url = "https://files.pythonhosted.org/packages/3c/4e/9eb2af5335545f9fbcd7af57bcf87c6025d31eaa31b14ec184a6c8675328/lxml-6.1.1-cp314-cp314t-win_amd64.whl", hash = "sha256:793033d6c5cdf33a573f910d9bea14ef8f5771820411d118da8e1182edb53d5e", size = 4393350, upload-time = "2026-05-18T19:18:10.076Z" },
+    { url = "https://files.pythonhosted.org/packages/7f/2c/0f1e93c636720e8a3eb59af2bfda99d98b55891e1c53bc30c2e0e865f01b/lxml-6.1.1-cp314-cp314t-win_arm64.whl", hash = "sha256:58bb955caba94e467d2a96da17660d2d704e0675894cba21ab8a775b8621fd1c", size = 3817223, upload-time = "2026-05-19T19:22:56.823Z" },
+]
+
 [[package]]
 name = "markdown-it-py"
 version = "4.2.0"
@@ -791,6 +871,12 @@ dependencies = [
 dev = [
     { name = "ruff" },
 ]
+web2md = [
+    { name = "beautifulsoup4" },
+    { name = "httpx" },
+    { name = "lxml" },
+    { name = "markdownify" },
+]
 
 [package.metadata]
 requires-dist = [
@@ -800,6 +886,12 @@ requires-dist = [
 
 [package.metadata.requires-dev]
 dev = [{ name = "ruff", specifier = ">=0.8" }]
+web2md = [
+    { name = "beautifulsoup4", specifier = ">=4.15.0" },
+    { name = "httpx", specifier = ">=0.28.1" },
+    { name = "lxml", specifier = ">=6.1.1" },
+    { name = "markdownify", specifier = ">=1.2.3" },
+]
 
 [[package]]
 name = "mdurl"
diff --git a/web2md/README.md b/web2md/README.md
new file mode 100644
index 0000000..adb09ff
--- /dev/null
+++ b/web2md/README.md
@@ -0,0 +1,39 @@
+# web2md
+
+Sometimes clean, structured Markdown files are not available, so they must be
+fetched from a website and converted. This directory holds a deterministic
+scraper for the [Google developer documentation style
+guide](https://developers.google.com/style). The output is one file under
+`md/`, ready for `make wiki-sandbox`.
+
+The result is a dated snapshot of a living document. Re-run with `--refresh` to
+update it.
+
+## Fetch and convert
+
+```bash
+# Install scraper deps (separate from the heavy marker-pdf stack)
+uv sync --group web2md
+
+# Fetch (or reuse web2md/cache/) and write md/GoogleDeveloperDocumentationStyleGuide.md
+make style-guide
+
+# Or call the module directly:
+uv run --group web2md python web2md/google_style_guide.py
+uv run --group web2md python web2md/google_style_guide.py --refresh
+```
+
+## Markdown linting
+
+The generated file under `md/` can be checked manually with the same tools as
+`pdf2md`:
+
+```bash
+prettier --check md/GoogleDeveloperDocumentationStyleGuide.md
+prettier --write md/GoogleDeveloperDocumentationStyleGuide.md  # Edits in place!
+
+markdownlint-cli2 md/GoogleDeveloperDocumentationStyleGuide.md
+markdownlint-cli2 --fix md/GoogleDeveloperDocumentationStyleGuide.md  # Edits in place!
+
+cspell md/GoogleDeveloperDocumentationStyleGuide.md
+```
diff --git a/web2md/google_style_guide.py b/web2md/google_style_guide.py
new file mode 100644
index 0000000..73bd108
--- /dev/null
+++ b/web2md/google_style_guide.py
@@ -0,0 +1,668 @@
+"""Fetch the Google developer documentation style guide into one Markdown file."""
+
+from __future__ import annotations
+
+import argparse
+import re
+import sys
+import time
+import warnings
+from dataclasses import dataclass
+from datetime import UTC, datetime
+from html import unescape
+from pathlib import Path
+from urllib.parse import urljoin, urlparse
+
+import httpx
+from bs4 import BeautifulSoup, NavigableString, Tag
+from markdownify import ATX, MarkdownConverter
+
+BASE = "https://developers.google.com"
+BOOK = f"{BASE}/style"
+LOCALE = {"hl": "en"}
+UA = "md2okf-web2md/0.1 (+https://github.com/lars20070/md2okf)"
+NAV = ".devsite-book-nav-wrapper"
+BODY = "div.devsite-article-body"
+DROP = (
+    "devsite-key-takeaways-panel",
+    "devsite-recommendations",
+    "devsite-feedback",
+    "devsite-thumb-rating",
+    "devsite-toc",
+    ".devsite-rating-container",
+    ".devsite-content-footer",
+    ".material-icons",
+    "[aria-hidden=true]",
+    "script",
+    "style",
+    "noscript",
+    "iframe",
+)
+ICON_TEXT = {
+    "icon-dontuse": "Don't use:",
+    "icon-avoid": "Avoid:",
+    "icon-android": "(Android)",
+    "icon-cloud": "(Cloud)",
+    "icon-workspace": "(Workspace)",
+}
+ASIDE_KIND = {
+    "note": "NOTE",
+    "caution": "CAUTION",
+    "warning": "WARNING",
+    "success": "TIP",
+}
+REQUEST_DELAY_S = 0.2
+MAX_RETRIES = 5
+WORD_LIST_TERM_EXPECTED = 598
+WORD_LIST_TERM_TOLERANCE = 30
+SIZE_MIN = 0.4 * 1024 * 1024
+SIZE_MAX = 1.2 * 1024 * 1024
+
+ROOT = Path(__file__).resolve().parent
+DEFAULT_CACHE = ROOT / "cache"
+DEFAULT_OUTPUT = ROOT.parent / "md" / "GoogleDeveloperDocumentationStyleGuide.md"
+
+_SANITIZE_RE = re.compile(r"[^A-Za-z0-9._-]")
+_MD_LINK_RE = re.compile(r"\[([^\]]*)\]\((#[^)]+)\)")
+_HEADING_RE = re.compile(r"^(#{1,6})\s+", re.MULTILINE)
+_PANDOC_DD_RE = re.compile(r"^:\s{3,}", re.MULTILINE)
+_DEVSITE_RE = re.compile(r"devsite-|material-icons", re.IGNORECASE)
+
+
+@dataclass(frozen=True)
+class Page:
+    section: str
+    title: str
+    slug: str
+    url: str
+    path: str
+
+
+def slug_from_path(path: str) -> str:
+    rest = path.removeprefix("/style").strip("/")
+    return rest or "style"
+
+
+def sanitize_id(raw: str) -> str:
+    def repl(match: re.Match[str]) -> str:
+        return f"u{ord(match.group(0)):04x}"
+
+    cleaned = _SANITIZE_RE.sub(repl, raw)
+    return cleaned or "id"
+
+
+def namespaced_anchor(page_slug: str, raw_id: str) -> str:
+    return f"{page_slug}--{sanitize_id(raw_id)}"
+
+
+def nav_text(el: Tag) -> str:
+    span = el.select_one(".devsite-nav-text")
+    text = span.get_text(" ", strip=True) if span else el.get_text(" ", strip=True)
+    return unescape(text)
+
+
+def discover_pages(html: str) -> list[Page]:
+    soup = BeautifulSoup(html, "lxml")
+    nav = soup.select_one(NAV)
+    if nav is None:
+        raise SystemExit(f"navigation not found: selector {NAV!r}")
+
+    pages: list[Page] = []
+    seen: set[str] = set()
+    sections: list[str] = []
+    current_section: str | None = None
+
+    for li in nav.find_all("li", recursive=True):
+        classes = li.get("class") or []
+        if "devsite-nav-heading" in classes:
+            current_section = nav_text(li) or current_section
+            if current_section and current_section not in sections:
+                sections.append(current_section)
+            continue
+        if current_section is None:
+            # Skip book-picker tabs (e.g. "Guides") that sit above the outline.
+            continue
+        if "devsite-nav-item" not in classes:
+            continue
+        link = li.find("a", href=True)
+        if link is None:
+            continue
+        href = link["href"].split("?")[0].split("#")[0]
+        if not href.startswith("/style"):
+            continue
+        full = urljoin(BASE, href)
+        if full in seen:
+            continue
+        seen.add(full)
+        path = urlparse(full).path.rstrip("/") or "/style"
+        pages.append(
+            Page(
+                section=current_section,
+                title=nav_text(link),
+                slug=slug_from_path(path),
+                url=full,
+                path=path,
+            )
+        )
+
+    if not (60 <= len(pages) <= 200):
+        raise SystemExit(f"unexpected page count: {len(pages)} (want 60–200)")
+    if len(sections) < 5:
+        raise SystemExit(f"unexpected section count: {len(sections)} (want ≥5)")
+    print(f"Discovered {len(pages)} pages in {len(sections)} sections.", file=sys.stderr)
+    return pages
+
+
+def fetch_html(
+    client: httpx.Client,
+    url: str,
+    cache_path: Path,
+    *,
+    refresh: bool,
+) -> str:
+    if cache_path.exists() and not refresh:
+        return cache_path.read_text(encoding="utf-8")
+
+    delay = REQUEST_DELAY_S
+    last_error: Exception | None = None
+    for attempt in range(MAX_RETRIES):
+        try:
+            response = client.get(url, params=LOCALE)
+            if response.status_code == 429 or response.status_code >= 500:
+                retry_after = response.headers.get("Retry-After")
+                wait = float(retry_after) if retry_after and retry_after.isdigit() else delay
+                print(
+                    f"  retry {attempt + 1}/{MAX_RETRIES} after HTTP {response.status_code}"
+                    f" ({wait:.1f}s): {url}",
+                    file=sys.stderr,
+                )
+                time.sleep(wait)
+                delay = min(delay * 2, 30.0)
+                continue
+            response.raise_for_status()
+            text = response.text
+            cache_path.parent.mkdir(parents=True, exist_ok=True)
+            cache_path.write_text(text, encoding="utf-8")
+            time.sleep(REQUEST_DELAY_S)
+            return text
+        except httpx.HTTPError as exc:
+            last_error = exc
+            print(
+                f"  retry {attempt + 1}/{MAX_RETRIES} after {exc!r} ({delay:.1f}s): {url}",
+                file=sys.stderr,
+            )
+            time.sleep(delay)
+            delay = min(delay * 2, 30.0)
+
+    raise SystemExit(f"failed to fetch {url}: {last_error}")
+
+
+def replace_icon_spans(body: Tag) -> None:
+    # Only rewrite icons inside definition terms. The word-list legend uses the
+    # same empty spans next to prose that already explains them.
+    for class_name, label in ICON_TEXT.items():
+        for el in body.select(f".{class_name}"):
+            if el.find_parent("dt") is None:
+                continue
+            el.clear()
+            el.append(NavigableString(f"{label} "))
+
+
+def unwrap_devsite_code(body: Tag) -> None:
+    for el in body.find_all("devsite-code"):
+        pre = el.find("pre")
+        if pre is not None:
+            el.replace_with(pre.extract())
+        else:
+            el.unwrap()
+
+
+def drop_noise(body: Tag, dropped_classes: set[str]) -> None:
+    for selector in DROP:
+        for el in body.select(selector):
+            for cls in el.get("class") or []:
+                dropped_classes.add(cls)
+            el.decompose()
+
+    # Permalink icons leave empty <a href="#…"> shells once material-icons go.
+    for a in list(body.find_all("a", href=True)):
+        if not a.get_text(strip=True) and not a.find("img"):
+            a.decompose()
+
+
+def clean_body(html: str, page: Page, dropped_classes: set[str]) -> Tag:
+    soup = BeautifulSoup(html, "lxml")
+    body = soup.select_one(BODY)
+    if body is None:
+        raise SystemExit(f"missing article body on {page.url} (selector {BODY!r})")
+
+    replace_icon_spans(body)
+    unwrap_devsite_code(body)
+    drop_noise(body, dropped_classes)
+
+    # Heading depth check before conversion (body uses h2–h4).
+    for heading in body.find_all(re.compile(r"^h[1-6]$")):
+        level = int(heading.name[1])
+        if level > 4:
+            raise SystemExit(f"body heading deeper than h4 on {page.url}: <{heading.name}>")
+        if level < 2 and heading.get("id") != "key-takeaways-panel-title":
+            # h1 should not appear in the article body for this book.
+            pass
+
+    return body
+
+
+def collect_ids(body: Tag, page_slug: str, anchor_map: dict[tuple[str, str], str]) -> None:
+    for el in body.find_all(attrs={"id": True}):
+        raw_id = el["id"]
+        if not raw_id or raw_id == "key-takeaways-panel-title":
+            continue
+        key = (page_slug, raw_id)
+        anchor_map[key] = namespaced_anchor(page_slug, raw_id)
+
+
+def absolutize_url(href: str) -> str:
+    return urljoin(BASE, href)
+
+
+def rewrite_internal_href(
+    href: str,
+    page: Page,
+    pages_by_path: dict[str, Page],
+    anchor_map: dict[tuple[str, str], str],
+    unresolved: list[str],
+) -> str:
+    parsed = urlparse(href)
+    fragment = parsed.fragment
+
+    # Same-page fragment.
+    if (not parsed.scheme and not parsed.netloc and not parsed.path) or href.startswith("#"):
+        if not fragment:
+            return href
+        key = (page.slug, fragment)
+        if key in anchor_map:
+            return f"#{anchor_map[key]}"
+        unresolved.append(f"{page.url} -> {href}")
+        return absolutize_url(f"{page.path}#{fragment}")
+
+    # Absolute or site-relative /style/... links.
+    if parsed.netloc and parsed.netloc != "developers.google.com":
+        return href
+    if parsed.scheme and parsed.scheme not in ("http", "https"):
+        return href
+
+    path = parsed.path.rstrip("/") or "/"
+    if not path.startswith("/style"):
+        if href.startswith("/") or parsed.netloc == "developers.google.com":
+            return absolutize_url(href)
+        return href
+
+    target_path = path if path != "/style" else "/style"
+    target_page = pages_by_path.get(target_path)
+    if target_page is None and target_path == "/style":
+        target_page = pages_by_path.get("/style")
+
+    if fragment:
+        if target_page is None:
+            unresolved.append(f"{page.url} -> {href}")
+            return absolutize_url(href)
+        key = (target_page.slug, fragment)
+        if key in anchor_map:
+            return f"#{anchor_map[key]}"
+        unresolved.append(f"{page.url} -> {href}")
+        return absolutize_url(href)
+
+    if target_page is not None:
+        return f"#{target_page.slug}"
+    unresolved.append(f"{page.url} -> {href}")
+    return absolutize_url(href)
+
+
+def apply_anchors_and_links(
+    body: Tag,
+    page: Page,
+    pages_by_path: dict[str, Page],
+    anchor_map: dict[tuple[str, str], str],
+    unresolved: list[str],
+) -> None:
+    # Inject an explicit anchor before every id-bearing element (sections,
+    # headings, dt terms, …). markdownify drops ids on non-heading tags, and
+    # many DevSite fragments live on <section id="…"> wrappers. DevSite often
+    # repeats the same id on a section and its heading — emit it once.
+    soup = body
+    while soup.parent is not None:
+        soup = soup.parent
+    emitted: set[str] = set()
+    for el in list(body.find_all(attrs={"id": True})):
+        raw_id = el["id"]
+        key = (page.slug, raw_id)
+        if key not in anchor_map:
+            continue
+        namespaced = anchor_map[key]
+        if namespaced not in emitted:
+            anchor_tag = soup.new_tag("a", attrs={"id": namespaced})
+            el.insert_before(anchor_tag)
+            emitted.add(namespaced)
+        del el["id"]
+
+    for a in body.find_all("a", href=True):
+        a["href"] = rewrite_internal_href(
+            a["href"], page, pages_by_path, anchor_map, unresolved
+        )
+
+    for img in body.find_all("img", src=True):
+        img["src"] = absolutize_url(img["src"])
+
+
+class StyleGuideConverter(MarkdownConverter):
+    """DevSite-specific HTML → Markdown rules."""
+
+    class Options(MarkdownConverter.Options):
+        heading_style = ATX
+        bullets = "-"
+        escape_asterisks = False
+        escape_underscores = False
+
+    def __init__(self, page_slug: str, aside_warnings: list[str], **options):
+        super().__init__(**options)
+        self.page_slug = page_slug
+        self.aside_warnings = aside_warnings
+
+    def convert_aside(self, el: Tag, text: str, parent_tags: set[str]) -> str:
+        classes = el.get("class") or []
+        kind = "NOTE"
+        matched = False
+        for cls, mapped in ASIDE_KIND.items():
+            if cls in classes:
+                kind = mapped
+                matched = True
+                break
+        if not matched and classes:
+            self.aside_warnings.append(f"unknown aside class {classes!r} on {self.page_slug}")
+        lines = [line for line in text.strip().splitlines() if line.strip()]
+        if not lines:
+            return "\n\n"
+        quoted = "\n> ".join(lines)
+        return f"\n\n> [!{kind}]\n> {quoted}\n\n"
+
+    def convert_a(self, el: Tag, text: str, parent_tags: set[str]) -> str:
+        # Keep explicit fragment targets; markdownify would otherwise drop empty
+        # anchors that have an id and no href.
+        anchor_id = el.get("id")
+        if anchor_id and not el.get("href") and not (text or "").strip():
+            return f'<a id="{anchor_id}"></a>'
+        return super().convert_a(el, text, parent_tags)
+
+    def process_text(self, el, parent_tags=None):
+        text = super().process_text(el, parent_tags=parent_tags)
+        # DevSite docs mention Markdown fences as literal ``` in prose; left
+        # alone those open a fence and swallow the following content.
+        if parent_tags is None or "pre" not in parent_tags:
+            text = text.replace("```", r"\`\`\`")
+        return text
+
+    def convert_dt(self, el: Tag, text: str, parent_tags: set[str]) -> str:
+        text = re.sub(r"\s+", " ", (text or "").strip())
+        if "_inline" in parent_tags:
+            return f" {text} "
+        if not text:
+            return "\n"
+        return f"\n\n**{text}**\n"
+
+    def convert_dd(self, el: Tag, text: str, parent_tags: set[str]) -> str:
+        text = (text or "").strip()
+        if "_inline" in parent_tags:
+            return f" {text} "
+        if not text:
+            return "\n"
+        indented = "\n".join(
+            f"    {line}" if line.strip() else "" for line in text.splitlines()
+        )
+        return f"\n{indented}\n"
+
+    def convert_hN(self, n: int, el: Tag, text: str, parent_tags: set[str]) -> str:
+        if "_inline" in parent_tags:
+            return text
+        # Document uses # / ## / ### for title / section / page, so body shifts +2.
+        n = max(1, min(6, n + 2))
+        text = re.sub(r"\s+", " ", (text or "").strip())
+        if not text:
+            return "\n\n"
+        return f"\n\n{'#' * n} {text}\n\n"
+
+    def convert_img(self, el: Tag, text: str, parent_tags: set[str]) -> str:
+        alt = el.get("alt") or ""
+        src = el.get("src") or ""
+        if not src:
+            return alt
+        return f"![{alt}]({src})"
+
+    def convert_pre(self, el: Tag, text: str, parent_tags: set[str]) -> str:
+        code = el.get_text().removeprefix("\n")
+        if not code.endswith("\n"):
+            code = code + "\n"
+        return f"\n\n```\n{code}```\n\n"
+
+
+def convert_body(body: Tag, page_slug: str, aside_warnings: list[str]) -> str:
+    converter = StyleGuideConverter(page_slug=page_slug, aside_warnings=aside_warnings)
+    return converter.convert(str(body)).strip()
+
+
+def build_toc(pages: list[Page]) -> str:
+    lines = ["## Table of contents", ""]
+    current_section = None
+    for page in pages:
+        if page.section != current_section:
+            if current_section is not None:
+                lines.append("")
+            current_section = page.section
+            lines.append(f"### {current_section}")
+            lines.append("")
+        lines.append(f"- [{page.title}](#{page.slug})")
+    lines.append("")
+    return "\n".join(lines)
+
+
+def assemble(pages: list[Page], bodies: dict[str, str], generated_at: str) -> str:
+    parts: list[str] = [
+        "---",
+        "title: Google Developer Documentation Style Guide",
+        f"source: {BOOK}",
+        f"generated_at: {generated_at}",
+        f"pages: {len(pages)}",
+        "format: Single-file Markdown",
+        "---",
+        "",
+        "# Google Developer Documentation Style Guide",
+        "",
+        f"*Snapshot of [{BOOK}]({BOOK}) generated {generated_at}.*",
+        "",
+        build_toc(pages),
+    ]
+
+    current_section = None
+    for page in pages:
+        if page.section != current_section:
+            current_section = page.section
+            parts.append(f"## {current_section}")
+            parts.append("")
+        parts.append(f'<a id="{page.slug}"></a>')
+        parts.append("")
+        parts.append(f"### {page.title}")
+        parts.append("")
+        parts.append(f"*Source: <{page.url}>*")
+        parts.append("")
+        parts.append(bodies[page.slug])
+        parts.append("")
+        parts.append("---")
+        parts.append("")
+
+    # Drop trailing divider.
+    while parts and parts[-1] in ("", "---"):
+        parts.pop()
+    parts.append("")
+    return "\n".join(parts)
+
+
+def markdown_without_fences(md: str) -> str:
+    """Strip fenced code blocks so example HTML does not trip content checks."""
+    return re.sub(r"```.*?```", "", md, flags=re.DOTALL)
+
+
+def validate_output(md: str, pages: list[Page], bodies: dict[str, str]) -> None:
+    errors: list[str] = []
+
+    if len(bodies) != len(pages):
+        missing = [p.slug for p in pages if p.slug not in bodies]
+        errors.append(f"pages written ({len(bodies)}) != discovered ({len(pages)}): {missing}")
+
+    emitted_anchors = set(re.findall(r'<a id="([^"]+)"></a>', md))
+    emitted_anchors.update(p.slug for p in pages)
+    for match in re.finditer(r'id="([^"]+)"', md):
+        emitted_anchors.add(match.group(1))
+
+    broken: list[str] = []
+    for label, target in _MD_LINK_RE.findall(md):
+        anchor = target[1:]  # strip #
+        if anchor not in emitted_anchors:
+            broken.append(f"[{label}]({target})")
+    if broken:
+        sample = "; ".join(broken[:20])
+        more = f" (+{len(broken) - 20} more)" if len(broken) > 20 else ""
+        errors.append(f"{len(broken)} broken internal links: {sample}{more}")
+
+    prose = markdown_without_fences(md)
+    if _DEVSITE_RE.search(prose):
+        errors.append("output still contains 'devsite-' or 'material-icons' outside code fences")
+    if _PANDOC_DD_RE.search(md):
+        errors.append("output still contains Pandoc-style definition list markers")
+
+    deep = [m.group(0).strip() for m in _HEADING_RE.finditer(md) if len(m.group(1)) > 6]
+    if deep:
+        errors.append(f"headings deeper than h6 found: {deep[:5]}")
+
+    size = len(md.encode("utf-8"))
+    if not (SIZE_MIN <= size <= SIZE_MAX):
+        errors.append(
+            f"output size {size / 1024 / 1024:.2f} MB outside "
+            f"{SIZE_MIN / 1024 / 1024:.1f}–{SIZE_MAX / 1024 / 1024:.1f} MB"
+        )
+
+    word_list = bodies.get("word-list", "")
+    # Definition terms are emitted as <a id="word-list--…"></a> then **term**.
+    term_count = len(
+        re.findall(r'<a id="word-list--[^"]+"></a>\s*\*\*', word_list)
+    )
+    low = WORD_LIST_TERM_EXPECTED - WORD_LIST_TERM_TOLERANCE
+    high = WORD_LIST_TERM_EXPECTED + WORD_LIST_TERM_TOLERANCE
+    if not (low <= term_count <= high):
+        errors.append(
+            f"word-list term count {term_count} outside {low}–{high} "
+            f"(expected ~{WORD_LIST_TERM_EXPECTED})"
+        )
+
+    if errors:
+        for err in errors:
+            print(f"VALIDATION ERROR: {err}", file=sys.stderr)
+        raise SystemExit(1)
+
+    print(
+        f"Validation OK: {len(pages)} pages, {size / 1024 / 1024:.2f} MB, "
+        f"{term_count} word-list terms, {len(emitted_anchors)} anchors.",
+        file=sys.stderr,
+    )
+
+
+def run(*, refresh: bool, cache_dir: Path, output: Path) -> None:
+    warnings.filterwarnings("ignore", category=UserWarning, module="bs4")
+    cache_dir.mkdir(parents=True, exist_ok=True)
+    output.parent.mkdir(parents=True, exist_ok=True)
+
+    with httpx.Client(
+        headers={"User-Agent": UA},
+        follow_redirects=True,
+        timeout=60.0,
+    ) as client:
+        index_html = fetch_html(
+            client, BOOK, cache_dir / "_index.html", refresh=refresh
+        )
+        pages = discover_pages(index_html)
+        pages_by_path = {p.path: p for p in pages}
+
+        cleaned: dict[str, Tag] = {}
+        dropped_classes: set[str] = set()
+        for i, page in enumerate(pages, start=1):
+            print(f"[{i}/{len(pages)}] {page.url}", file=sys.stderr)
+            html = fetch_html(
+                client, page.url, cache_dir / f"{page.slug}.html", refresh=refresh
+            )
+            cleaned[page.slug] = clean_body(html, page, dropped_classes)
+
+        if dropped_classes:
+            print(
+                "Dropped class values seen during cleaning: "
+                + ", ".join(sorted(dropped_classes)),
+                file=sys.stderr,
+            )
+
+        anchor_map: dict[tuple[str, str], str] = {}
+        for page in pages:
+            collect_ids(cleaned[page.slug], page.slug, anchor_map)
+
+        unresolved: list[str] = []
+        aside_warnings: list[str] = []
+        bodies: dict[str, str] = {}
+        for page in pages:
+            body = cleaned[page.slug]
+            apply_anchors_and_links(
+                body, page, pages_by_path, anchor_map, unresolved
+            )
+            bodies[page.slug] = convert_body(body, page.slug, aside_warnings)
+
+        if unresolved:
+            print(
+                f"{len(unresolved)} internal targets rewritten to absolute URLs "
+                f"(missing from map). Sample: {unresolved[:10]}",
+                file=sys.stderr,
+            )
+        for warning in aside_warnings:
+            print(f"WARNING: {warning}", file=sys.stderr)
+
+        md = assemble(
+            pages,
+            bodies,
+            generated_at=datetime.now(tz=UTC).date().isoformat(),
+        )
+        validate_output(md, pages, bodies)
+        output.write_text(md, encoding="utf-8")
+        print(f"Wrote {output}", file=sys.stderr)
+
+
+def main(argv: list[str] | None = None) -> None:
+    parser = argparse.ArgumentParser(
+        description="Fetch developers.google.com/style into one Markdown file."
+    )
+    parser.add_argument(
+        "--refresh",
+        action="store_true",
+        help="re-fetch pages even if a cache entry exists",
+    )
+    parser.add_argument(
+        "--cache-dir",
+        type=Path,
+        default=DEFAULT_CACHE,
+        help=f"HTML cache directory (default: {DEFAULT_CACHE})",
+    )
+    parser.add_argument(
+        "--output",
+        type=Path,
+        default=DEFAULT_OUTPUT,
+        help=f"output Markdown path (default: {DEFAULT_OUTPUT})",
+    )
+    args = parser.parse_args(argv)
+    run(refresh=args.refresh, cache_dir=args.cache_dir, output=args.output)
+
+
+if __name__ == "__main__":
+    main()

From 3bce28c041a49767609d01686b68b285a57f4f53 Mon Sep 17 00:00:00 2001
From: lars20070 <l.nilse@dunelm.org.uk>
Date: Fri, 31 Jul 2026 11:21:59 +0200
Subject: [PATCH 06/15] Fix frontmatter

---
 md/GoogleDeveloperDocumentationStyleGuide.md | 11 ++++++-----
 web2md/google_style_guide.py                 | 17 +++++++++--------
 2 files changed, 15 insertions(+), 13 deletions(-)

diff --git a/md/GoogleDeveloperDocumentationStyleGuide.md b/md/GoogleDeveloperDocumentationStyleGuide.md
index fba07ba..7656099 100644
--- a/md/GoogleDeveloperDocumentationStyleGuide.md
+++ b/md/GoogleDeveloperDocumentationStyleGuide.md
@@ -1,9 +1,10 @@
 ---
-title: Google Developer Documentation Style Guide
-source: https://developers.google.com/style
-generated_at: 2026-07-31
-pages: 70
-format: Single-file Markdown
+type: Book
+title: "Google. Google Developer Documentation Style Guide."
+description: "Style guide for Google developer documentation"
+resource: https://developers.google.com/style
+tags: [guide, Google]
+timestamp: 2026-07-31T09:20:57Z
 ---
 
 # Google Developer Documentation Style Guide
diff --git a/web2md/google_style_guide.py b/web2md/google_style_guide.py
index 73bd108..2c40b5f 100644
--- a/web2md/google_style_guide.py
+++ b/web2md/google_style_guide.py
@@ -464,19 +464,20 @@ def build_toc(pages: list[Page]) -> str:
     return "\n".join(lines)
 
 
-def assemble(pages: list[Page], bodies: dict[str, str], generated_at: str) -> str:
+def assemble(pages: list[Page], bodies: dict[str, str], timestamp: str) -> str:
     parts: list[str] = [
         "---",
-        "title: Google Developer Documentation Style Guide",
-        f"source: {BOOK}",
-        f"generated_at: {generated_at}",
-        f"pages: {len(pages)}",
-        "format: Single-file Markdown",
+        "type: Book",
+        'title: "Google. Google Developer Documentation Style Guide."',
+        'description: "Style guide for Google developer documentation"',
+        f"resource: {BOOK}",
+        "tags: [guide, Google]",
+        f"timestamp: {timestamp}",
         "---",
         "",
         "# Google Developer Documentation Style Guide",
         "",
-        f"*Snapshot of [{BOOK}]({BOOK}) generated {generated_at}.*",
+        f"*Snapshot of [{BOOK}]({BOOK}) generated {timestamp[:10]}.*",
         "",
         build_toc(pages),
     ]
@@ -632,7 +633,7 @@ def run(*, refresh: bool, cache_dir: Path, output: Path) -> None:
         md = assemble(
             pages,
             bodies,
-            generated_at=datetime.now(tz=UTC).date().isoformat(),
+            timestamp=datetime.now(tz=UTC).strftime("%Y-%m-%dT%H:%M:%SZ"),
         )
         validate_output(md, pages, bodies)
         output.write_text(md, encoding="utf-8")

From e6c8c953ba5a6037dee2e2dbb3c36e78d01ec9ce Mon Sep 17 00:00:00 2001
From: lars20070 <l.nilse@dunelm.org.uk>
Date: Sat, 1 Aug 2026 07:54:21 +0200
Subject: [PATCH 07/15] Keep plans folders

---
 .claude/plans/.gitkeep | 0
 .cursor/plans/.gitkeep | 0
 2 files changed, 0 insertions(+), 0 deletions(-)
 create mode 100644 .claude/plans/.gitkeep
 create mode 100644 .cursor/plans/.gitkeep

diff --git a/.claude/plans/.gitkeep b/.claude/plans/.gitkeep
new file mode 100644
index 0000000..e69de29
diff --git a/.cursor/plans/.gitkeep b/.cursor/plans/.gitkeep
new file mode 100644
index 0000000..e69de29

From 9823e9692e36a77a1dea47b3a7a3bb2d8878e134 Mon Sep 17 00:00:00 2001
From: lars20070 <l.nilse@dunelm.org.uk>
Date: Sat, 1 Aug 2026 08:00:28 +0200
Subject: [PATCH 08/15] Fix Google style guide with markdownlint-cli2

---
 md/GoogleDeveloperDocumentationStyleGuide.md | 65 +++++++++++---------
 1 file changed, 35 insertions(+), 30 deletions(-)

diff --git a/md/GoogleDeveloperDocumentationStyleGuide.md b/md/GoogleDeveloperDocumentationStyleGuide.md
index 7656099..8cd910b 100644
--- a/md/GoogleDeveloperDocumentationStyleGuide.md
+++ b/md/GoogleDeveloperDocumentationStyleGuide.md
@@ -2523,6 +2523,7 @@ This guide is *not* intended to do the following:
 
 > [!NOTE]
 > **Note**: Two disclaimers:
+>
 > - The guidance in this style guide doesn't limit the changes that Google can
 >   make to its documentation.
 > - If you don't read a given guideline, then you are still responsible for
@@ -7444,6 +7445,7 @@ unambiguous.
   | If the VM has started and if you're able to connect... | If the VM has started and you're able to connect... |
   | The resource hierarchy design creates both IAM segmentation and network segmentation by default. | The resource hierarchy design creates both IAM and network segmentation by default. |
   | An egress rule whose action is `allow`, whose destination is `0.0.0.0/0`, and whose priority is the lowest possible (`65535`). | An egress rule whose action is `allow`, destination is `0.0.0.0/0`, and priority is the lowest possible (`65535`). |
+
 - Use helper words. Helper words such as *then*, *that*, and *of*
   are frequently left out of conversational English. Use these words to avoid ambiguity.
 
@@ -10105,12 +10107,12 @@ Recommended:
 We use your feedback to improve the Animals API, in accordance with Example
 Pet Store's Privacy Policy:
 
-http://www.examplepetstore.com/privacy/
+<http://www.examplepetstore.com/privacy/>
 
 Not recommended:
 
 We use your feedback to improve the Animals API, in accordance with Example
-Pet Store's Privacy Policy at http://www.examplepetstore.com/privacy/.
+Pet Store's Privacy Policy at <http://www.examplepetstore.com/privacy/>.
 
 When you do put a period after a URL, don't leave any space between the last character of
 the URL and the period.
@@ -10416,8 +10418,6 @@ Not recommended: For example, a disaster
 relief map is not subject to the usage limits even if it has been
 developed/hosted by a commercial entity.
 
-  
-
 Recommended: Call this method five or six
 times.
 
@@ -10438,8 +10438,6 @@ and edit your own data.
 Not recommended: You can
 view and/or edit your own data.
 
-  
-
 Recommended: You can
 export raw events, processed events, or both.
 
@@ -10458,13 +10456,13 @@ Use forward slashes, as appropriate, in computer file paths and URLs.
 > **Note**: If you're documenting a Windows path, use backslashes.
 
 Recommended:
-https://developers.google.com/cardboard/
+<https://developers.google.com/cardboard/>
 
 Where very long URLs extend beyond a line, add a line break immediately after
 a slash. Don't ever insert an extraneous hyphen into a URL to break it between two lines.
 
 Recommended:
-https://developers.google.com/
+<https://developers.google.com/>
   
 cardboard/
 
@@ -10692,7 +10690,7 @@ examples using *such as*, *for example*, or *like* in various ways.
 | --- | --- | --- |
 | Introduce a short-to-medium-length example at the end of a sentence. For clarity, consider setting off the example using a comma, parentheses, or an em dash as appropriate. Avoid using a semicolon for this purpose. | Choose a strong encryption algorithm, such as AES-256.  You can monitor various metrics for your managed database instances—for example, CPU utilization, storage capacity, and active connections.  The API supports common image formats like PNG and JPEG. | Enter a name for the instance, for example, `my-instance-99`.  Specify the region for deployment; for example, `us-central1`.  Enter a name for the instance (for example, `my-instance-99`). |
 | Introduce a short example in the middle of a sentence. Keep the example in the middle of a sentence relatively short and consider setting it off with dashes, commas, or parentheses as appropriate. | Enter a six-digit hex number (for example, `228B22`), and then click **OK**.  The virtual machine (VM) requires an operating system, such as Ubuntu 22.04, to be installed.  Some elements, like buttons and input fields, have default accessibility attributes. | Enter a six-digit hex number (for example, if you want the color forest green, enter `228B22`), and then click **OK**. |
-| Introduce a longer example as a separate sentence. For a longer example, introduce it as a separate sentence using *for example* as an adverb in that sentence. | You can assign tags to your virtual machine instances to categorize them. For example, you could tag instances by environment with `env:prod` or `env:dev`. |  |
+| Introduce a longer example as a separate sentence. For a longer example, introduce it as a separate sentence using *for example* as an adverb in that sentence. | You can assign tags to your virtual machine instances to categorize them. For example, you could tag instances by environment with `env:prod` or `env:dev`. | |
 
 ---
 
@@ -10784,7 +10782,6 @@ ecommerce application:
 <a id="images--bounded"></a>
 ![Bounded contexts are applied to an application.](https://cloud.google.com/architecture/images/microservices-architecture-refactoring-monoliths-bounded-contexts.svg)
 
-
 **Figure 1**. Application capabilities are separated into bounded contexts that
 migrate to services.
 
@@ -11378,6 +11375,7 @@ that applies to standard text also applies to headings—for example,
 
   ### Estimate costs
   ```
+
 - **Don't use empty headings**. Make sure headings are followed by content.
 
   Recommended:
@@ -12128,20 +12126,14 @@ rules.</aside>
 > **Note**: All VPC networks include firewall
 > rules.
 
-
-
 > [!CAUTION]
 > **Caution**: We don't recommend using a subnet that's part of a dynamic
 > route.
 
-
-
 > [!WARNING]
 > **Warning**: Do not manually edit or delete generated
 > table entries.
 
-
-
 > [!TIP]
 > **Success**: You've successfully created a Compute Engine instance.
 
@@ -12913,10 +12905,10 @@ Recommended:
 | Guidance | Recommended | Not recommended |
 | --- | --- | --- |
 | Make sure that the first sentence in a procedural step includes an imperative verb. | Clone the repository that contains the sample data. | You need the project ID later in this document. Retrieve the project ID. |
-| Use complete sentences. |  |  |
+| Use complete sentences. | | |
 | Use parallel structure and consistent verb form. | Download the service account key to your local machine. Click **More**, and then click **Download**. | Download the service account key to your local machine by clicking **More** and then clicking **Download** file. |
 | For an optional step, type *Optional:* as the first word of the step. | Optional: Type an arbitrary string... | (Optional) Type an arbitrary string... |
-| Set the context (such as a tool or an environment) in which the reader performs a procedure.  If there are multiple headings associated with a set of procedures, restate the context of the procedure in the first step, even if the context is the same as in the previous procedure. | In Cloud Shell, connect to the development cluster.  In the Google Cloud console, go to the **BigQuery** page. |  |
+| Set the context (such as a tool or an environment) in which the reader performs a procedure.  If there are multiple headings associated with a set of procedures, restate the context of the procedure in the first step, even if the context is the same as in the previous procedure. | In Cloud Shell, connect to the development cluster.  In the Google Cloud console, go to the **BigQuery** page. | |
 | Write in the order that the reader needs to follow. State the location of the action before stating the action. | In Google Docs, click **File > New > Document**.  In the Google Cloud console, go to the **Monitoring** page. | Click **File > New > Document** in Google Docs.  Go to the **Monitoring** page in the Google Cloud console. |
 | State the purpose or goal of the action before stating the action. | To start a new document, click **File > New > Document**. | Click **File > New > Document** to start a new document. |
 | Don't use directional language to orient the reader, such as *above*, *below*, or *right-hand side*. This type of language doesn't work well for accessibility or for localization. If a UI element is hard to find, provide a screenshot.  For information about documenting icons, see [Buttons and icons](#ui-elements--buttons). | Click **Menu**.  In the preceding diagram,...  In the following diagram,... | Click the button with three lines.  In the above diagram, ...  In the diagram below, ... |
@@ -12924,12 +12916,12 @@ Recommended:
 | Avoid using *run the following command* to introduce code. Instead, focus on what the command does. | In Cloud Shell, deploy the load generator:...  Define a firewall rule to allow internal traffic:... | In Cloud Shell, deploy the load generator by running the following command:...  Run the following command:... |
 | If the reader must press **Enter** after a step, then include that instruction as part of the step. | Click the search box, type `custom function`, and then press **Enter**. | 1. Click the search box and type `custom function`. 2. Press **Enter**. |
 | Don't include keyboard shortcuts. | Copy the command, and then paste it... | Press Ctrl+C, and then press Ctrl+V... |
-| When there's more than one way to do something, give only the best way. Giving alternate ways can confuse readers. |  |  |
-| If your procedure includes code samples, see how to format  [code samples](#code-samples). |  |  |
-| If your procedure includes commands, see how to format  [commands](#code-syntax--formatting-a-command). |  |  |
-| Ensure that the reader has the information that they need in order to prepare for the task ahead of time. Having information in advance supports task management, executive functioning, memory, and emotional regulation. | The following hardware and software are required:... |  |
-| Include as few steps as possible to complete the task. Limit interruptions in the path. |  |  |
-| Focus on one reader decision at a time. Separate each instruction by making each instruction a separate list item. |  |  |
+| When there's more than one way to do something, give only the best way. Giving alternate ways can confuse readers. | | |
+| If your procedure includes code samples, see how to format  [code samples](#code-samples). | | |
+| If your procedure includes commands, see how to format  [commands](#code-syntax--formatting-a-command). | | |
+| Ensure that the reader has the information that they need in order to prepare for the task ahead of time. Having information in advance supports task management, executive functioning, memory, and emotional regulation. | The following hardware and software are required:... | |
+| Include as few steps as possible to complete the task. Limit interruptions in the path. | | |
+| Focus on one reader decision at a time. Separate each instruction by making each instruction a separate list item. | | |
 
 ---
 
@@ -13588,6 +13580,7 @@ then provide that context. The following are a few such situations:
   ```
     <a href="mailto:support@example.com">send email to Technical Support</a>
   ```
+
 - **Links to sections on the same page.** When you're
   linking to another section on the same page, let the reader know that the link
   takes you to a different section of the same page. Use a standard phrase to clue
@@ -14086,7 +14079,7 @@ list:
 | --- | --- |
 | Attribute names and values | The `imageURL` attribute contains the path for the image file that you can open in a browser—for example, `https://www.example.com/images/product.jpg`.  You can create a VM instance using the `e2-highcpu-16` machine type in the `us-central1-a` region. |
 | Class names | The `SnapshotDiskOperator` class includes the `generate_snapshot_name` method. |
-| Command output | The output is similar to the following:     ```         Found sysprep-specialize-script-ps1 in metadata.         ...         Finished running specialize scripts.          ``` |
+| Command output | The output is similar to the following:     ```Found sysprep-specialize-script-ps1 in metadata.         ...         Finished running specialize scripts.``` |
 | [Command-line utility names](https://developers.google.com/style/code-in-text#tool-names), such as `gcloud`, `gsutil`, `kubectl`, and `bq` | You can use the `kubectl` tool to define a network policy. |
 | Data types | Nested data is represented as a `STRUCT` type. |
 | Database elements (such as row and column names) | The query extracts the `month`, `julianday`, and `dayofweek` values from the `datetime` and `timestamp` columns. |
@@ -14129,8 +14122,8 @@ or as a code entity like an attribute or value, then use code font.
 | Item | Recommended |
 | --- | --- |
 | Domain names | The test environment is designed only for standard application offerings from example.com. |
-|<a id="code-in-text--u0023tool-names"></a> Names of products, services, and organizations | Example Organization has current and former employees who use Google products such as Google Docs and Google Sheets. |
-| URLs that the reader is supposed to follow in a browser | You can find support at https://support.example.com.  It's usually best to format a URL as a link and use descriptive link text instead of exposing the URL itself. For more information, see [Avoid URLs as link text](#cross-references--urls). |
+| <a id="code-in-text--u0023tool-names"></a> Names of products, services, and organizations | Example Organization has current and former employees who use Google products such as Google Docs and Google Sheets. |
+| URLs that the reader is supposed to follow in a browser | You can find support at <https://support.example.com>.  It's usually best to format a URL as a link and use descriptive link text instead of exposing the URL itself. For more information, see [Avoid URLs as link text](#cross-references--urls). |
 
 <a id="code-in-text--code-in-ui"></a>
 
@@ -14316,6 +14309,7 @@ To link to a class or method:
   ```
   <a href="/reference/android/widget/TextView">TextView</a>
   ```
+
 - To link to a method, use the method name as a fragment identifier. If
   you're linking to a static method, also include the class name in the link
   text. If you need to distinguish between overloaded versions of a particular
@@ -14324,6 +14318,7 @@ To link to a class or method:
   ```
   <a href="/reference/android/app/Activity.html#onCreate(android.os.Bundle)">onCreate(Bundle)</a>
   ```
+
 - To link the attribute for a particular widget or layout to its Javadoc
   in the widget or layout's API reference entry, use the URL for the page, and
   then add the fragment identifier
@@ -14492,6 +14487,7 @@ following best practices:
   ```
   gcloud compute ssh
   ```
+
 - **Determine which arguments are needed to complete each task in the recommended way**.
   To minimize the number of options that you need to document in non-reference content, use as
   few optional arguments as possible. Rely on the command reference for the complete list of
@@ -14570,7 +14566,7 @@ Recommended:
 Enter the following code into the terminal:
 
 ```
-$ adb devices
+adb devices
 ```
 
 The output is the following:
@@ -14601,7 +14597,7 @@ lines, we recommend using separate code blocks for input and output.
 Recommended:
 
 ```
-$ cat ~/.ssh/my-ssh-key.pub
+cat ~/.ssh/my-ssh-key.pub
 ```
 
 The output is similar to the following:
@@ -14735,6 +14731,7 @@ commands. Instead, choose one of the following approaches:
       --source-file=SOURCE_FILE \
       --byol
   ```
+
 - **Document optional arguments in separate tasks**. In some cases, it might be best to
   treat different options in separate sections.
 
@@ -14768,6 +14765,7 @@ commands. Instead, choose one of the following approaches:
       --source-file=SOURCE_FILE \
       --data-disk
   ```
+
 - **Let the reader know that the command contains optional arguments**. If you must
   include special characters to indicate optional arguments, indicate that fact when you
   introduce the command.
@@ -15044,6 +15042,7 @@ If your sample code and command placeholders occur in a sentence, use the follow
   ```
   <code><var>PLACEHOLDER_NAME</var></code>
   ```
+
 - In Markdown, wrap inline placeholders in backticks (`), and use an
   asterisk (*) before the first backtick and after the second one
   (`` *`PLACEHOLDER_NAME`* ``).
@@ -15076,12 +15075,15 @@ If your placeholders are in a block of code, use the following formatting:
         </pre>
         
   ```
+
 - In Markdown, wrap the code block in a code fence (\`\`\`). Inside a
   code fence, you can't apply formatting like bold or italic.
 
   ```
   ```
+
   PLACEHOLDER_NAME
+
   ```
   ```
 
@@ -15194,6 +15196,7 @@ Use the following format for two or more placeholders:
   ```
   <li><code><var>INSTANCE_NAME</var></code>: description</li>
   ```
+
 - If the description contains an example, introduce it with an *em dash* or
   *such as*—for example:
 
@@ -15204,6 +15207,7 @@ Use the following format for two or more placeholders:
   ```
   <li><code><var>INSTANCE_NAME</var></code>: description, such as...</li>
   ```
+
 - Each item in the list follows our [list style](#lists).
 
 Recommended:
@@ -15265,6 +15269,7 @@ sample output:
   ```
   <li><code><var>INSTANCE_NAME</var></code>: description</li>
   ```
+
 - If the description contains an example, introduce it with an *em dash* or
   *such as*—for example:
 
@@ -16131,7 +16136,7 @@ If you need a generic email address, use one of the domains listed
 in
 [Example domain names](#examples--example-domain-names)
 and one of the names listed in [Example person names](#examples--example-person-names)—for
-example, dana@example.com. It's OK to use generic addresses like support@example.net. Don't use
+example, <dana@example.com>. It's OK to use generic addresses like <support@example.net>. Don't use
 person names, product names, or made-up names in email addresses.
 
 <a id="examples--names"></a>

From 2ee0ec8a7f42d6fae9feb120f921dfad3bd6f732 Mon Sep 17 00:00:00 2001
From: lars20070 <l.nilse@dunelm.org.uk>
Date: Sat, 1 Aug 2026 08:14:29 +0200
Subject: [PATCH 09/15] Rename scraping script to web2md.py

---
 .../plans/google-style-guide-to-markdown.md   | 26 ++++++++++++-------
 Makefile                                      |  2 +-
 web2md/README.md                              |  4 +--
 3 files changed, 20 insertions(+), 12 deletions(-)

diff --git a/.cursor/plans/google-style-guide-to-markdown.md b/.cursor/plans/google-style-guide-to-markdown.md
index 344cbaa..3772d65 100644
--- a/.cursor/plans/google-style-guide-to-markdown.md
+++ b/.cursor/plans/google-style-guide-to-markdown.md
@@ -146,7 +146,7 @@ place to confirm it.
 
 ## Changes
 
-### `web2md/google_style_guide.py` (new)
+### `web2md/web2md.py` (new)
 
 One module, argparse CLI, six stages. DevSite-specific selectors live in
 constants at the top so pointing it at another DevSite book is a config change.
@@ -154,16 +154,24 @@ constants at the top so pointing it at another DevSite book is a config change.
 ```python
 BASE = "https://developers.google.com"
 BOOK = f"{BASE}/style"
-LOCALE = {"hl": "en"}          # pin the locale; the page honours ?hl=en
+LOCALE = {"hl": "en"}  # pin the locale; the page honours ?hl=en
 UA = "md2okf-web2md/0.1 (+https://github.com/<owner>/md2okf)"
 NAV = ".devsite-book-nav-wrapper"
 BODY = "div.devsite-article-body"
 DROP = (
-    "devsite-key-takeaways-panel", "devsite-recommendations",
-    "devsite-feedback", "devsite-thumb-rating", "devsite-toc",
-    ".devsite-rating-container", ".devsite-content-footer",
-    ".material-icons", "[aria-hidden=true]",
-    "script", "style", "noscript", "iframe",
+    "devsite-key-takeaways-panel",
+    "devsite-recommendations",
+    "devsite-feedback",
+    "devsite-thumb-rating",
+    "devsite-toc",
+    ".devsite-rating-container",
+    ".devsite-content-footer",
+    ".material-icons",
+    "[aria-hidden=true]",
+    "script",
+    "style",
+    "noscript",
+    "iframe",
 )
 ```
 
@@ -245,7 +253,7 @@ web2md = [
 ]
 ```
 
-Run it as `uv run --group web2md python web2md/google_style_guide.py`. `pyyaml`
+Run it as `uv run --group web2md python web2md/web2md.py`. `pyyaml`
 from the old plan is not needed.
 
 ### `Makefile`
@@ -256,7 +264,7 @@ from the old plan is not needed.
 ```make
 # Fetch the Google developer documentation style guide into md/ as one file.
 style-guide:
-	uv run --group web2md python web2md/google_style_guide.py
+	uv run --group web2md python web2md/web2md.py
 ```
 
 `ruff check .` already covers the new module.
diff --git a/Makefile b/Makefile
index ec70f5c..2ae014a 100644
--- a/Makefile
+++ b/Makefile
@@ -60,4 +60,4 @@ wiki-sandbox:
 
 # Fetch the Google developer documentation style guide into md/ as one file.
 style-guide:
-	uv run --group web2md python web2md/google_style_guide.py
+	uv run --group web2md python web2md/web2md.py
diff --git a/web2md/README.md b/web2md/README.md
index adb09ff..48f4c60 100644
--- a/web2md/README.md
+++ b/web2md/README.md
@@ -19,8 +19,8 @@ uv sync --group web2md
 make style-guide
 
 # Or call the module directly:
-uv run --group web2md python web2md/google_style_guide.py
-uv run --group web2md python web2md/google_style_guide.py --refresh
+uv run --group web2md python web2md/web2md.py
+uv run --group web2md python web2md/web2md.py --refresh
 ```
 
 ## Markdown linting

From b195e45c6c7817603cdc4e2e5cedb1ef1288c0a6 Mon Sep 17 00:00:00 2001
From: lars20070 <l.nilse@dunelm.org.uk>
Date: Sat, 1 Aug 2026 08:29:58 +0200
Subject: [PATCH 10/15] Rename make target

---
 Makefile                                     | 4 ++--
 md/GoogleDeveloperDocumentationStyleGuide.md | 6 +++---
 web2md/README.md                             | 4 ++--
 web2md/{google_style_guide.py => web2md.py}  | 2 +-
 4 files changed, 8 insertions(+), 8 deletions(-)
 rename web2md/{google_style_guide.py => web2md.py} (99%)

diff --git a/Makefile b/Makefile
index 2ae014a..0bfc026 100644
--- a/Makefile
+++ b/Makefile
@@ -58,6 +58,6 @@ wiki-container:
 wiki-sandbox:
 	./scripts/compile-wiki-sandbox.sh
 
-# Fetch the Google developer documentation style guide into md/ as one file.
-style-guide:
+# Fetch the website into md/ as one file.
+scrape:
 	uv run --group web2md python web2md/web2md.py
diff --git a/md/GoogleDeveloperDocumentationStyleGuide.md b/md/GoogleDeveloperDocumentationStyleGuide.md
index 8cd910b..0bc328d 100644
--- a/md/GoogleDeveloperDocumentationStyleGuide.md
+++ b/md/GoogleDeveloperDocumentationStyleGuide.md
@@ -1,15 +1,15 @@
 ---
-type: Book
+type: Website
 title: "Google. Google Developer Documentation Style Guide."
 description: "Style guide for Google developer documentation"
 resource: https://developers.google.com/style
 tags: [guide, Google]
-timestamp: 2026-07-31T09:20:57Z
+timestamp: 2026-08-01T06:28:39Z
 ---
 
 # Google Developer Documentation Style Guide
 
-*Snapshot of [https://developers.google.com/style](https://developers.google.com/style) generated 2026-07-31.*
+*Snapshot of [https://developers.google.com/style](https://developers.google.com/style) generated 2026-08-01.*
 
 ## Table of contents
 
diff --git a/web2md/README.md b/web2md/README.md
index 48f4c60..037f0bd 100644
--- a/web2md/README.md
+++ b/web2md/README.md
@@ -15,8 +15,8 @@ update it.
 # Install scraper deps (separate from the heavy marker-pdf stack)
 uv sync --group web2md
 
-# Fetch (or reuse web2md/cache/) and write md/GoogleDeveloperDocumentationStyleGuide.md
-make style-guide
+# Fetch (or reuse web2md/cache/) and write md/*.md
+make scrape
 
 # Or call the module directly:
 uv run --group web2md python web2md/web2md.py
diff --git a/web2md/google_style_guide.py b/web2md/web2md.py
similarity index 99%
rename from web2md/google_style_guide.py
rename to web2md/web2md.py
index 2c40b5f..d41ca6b 100644
--- a/web2md/google_style_guide.py
+++ b/web2md/web2md.py
@@ -467,7 +467,7 @@ def build_toc(pages: list[Page]) -> str:
 def assemble(pages: list[Page], bodies: dict[str, str], timestamp: str) -> str:
     parts: list[str] = [
         "---",
-        "type: Book",
+        "type: Website",
         'title: "Google. Google Developer Documentation Style Guide."',
         'description: "Style guide for Google developer documentation"',
         f"resource: {BOOK}",

From ec4032b1d612b174d05758871d2d19da577ad75d Mon Sep 17 00:00:00 2001
From: lars20070 <l.nilse@dunelm.org.uk>
Date: Sat, 1 Aug 2026 09:39:13 +0200
Subject: [PATCH 11/15] Unit tests for web scraper

---
 .coderabbit.yaml               |  26 ++++-
 .cspell.json                   |  15 ++-
 .github/workflows/ci.yml       |  24 ++++
 .gitignore                     |   4 +-
 AGENTS.md                      |  10 ++
 Makefile                       |  14 ++-
 README.md                      |  17 ++-
 pyproject.toml                 |  32 +++++-
 uv.lock                        |  38 +++++++
 web2md/README.md               |  29 ++++-
 web2md/{ => src}/web2md.py     | 120 +++++++++++++++++---
 web2md/tests/conftest.py       |  98 ++++++++++++++++
 web2md/tests/test_assemble.py  | 102 +++++++++++++++++
 web2md/tests/test_cleaning.py  | 182 ++++++++++++++++++++++++++++++
 web2md/tests/test_cli.py       |  65 +++++++++++
 web2md/tests/test_converter.py | 123 ++++++++++++++++++++
 web2md/tests/test_discovery.py |  63 +++++++++++
 web2md/tests/test_fetch.py     | 198 +++++++++++++++++++++++++++++++++
 web2md/tests/test_links.py     | 180 ++++++++++++++++++++++++++++++
 web2md/tests/test_run.py       | 147 ++++++++++++++++++++++++
 web2md/tests/test_slugs.py     |  70 ++++++++++++
 web2md/tests/test_validate.py  | 153 +++++++++++++++++++++++++
 22 files changed, 1680 insertions(+), 30 deletions(-)
 rename web2md/{ => src}/web2md.py (80%)
 create mode 100644 web2md/tests/conftest.py
 create mode 100644 web2md/tests/test_assemble.py
 create mode 100644 web2md/tests/test_cleaning.py
 create mode 100644 web2md/tests/test_cli.py
 create mode 100644 web2md/tests/test_converter.py
 create mode 100644 web2md/tests/test_discovery.py
 create mode 100644 web2md/tests/test_fetch.py
 create mode 100644 web2md/tests/test_links.py
 create mode 100644 web2md/tests/test_run.py
 create mode 100644 web2md/tests/test_slugs.py
 create mode 100644 web2md/tests/test_validate.py

diff --git a/.coderabbit.yaml b/.coderabbit.yaml
index e640099..cc2fe81 100644
--- a/.coderabbit.yaml
+++ b/.coderabbit.yaml
@@ -82,6 +82,25 @@ reviews:
         style: flag commands, paths, options, and examples that this change has
         made wrong or stale, and instructions that contradict each other. Do
         not raise wording, grammar, or line-length nitpicks.
+    - path: "**/*.py"
+      instructions: |
+        Review Python code for:
+        - PEP 8, type hints, docstrings, error handling, and performance. Flag mutable default arguments.
+        - Blocking calls inside async functions (use asyncio equivalents)
+        - API keys or secrets accidentally logged or exposed in error messages
+        - Missing context managers for network/file resources
+        - Unbounded retries or missing timeouts on external calls
+        - User-controlled strings passed unsanitized to shell commands or LLM prompts
+    - path: "**/tests/**/*.py"
+      instructions: |
+        Review test code for:
+        - Check pytest best practices, fixture usage, test isolation, and edge case coverage.
+        - Async test functions must have @pytest.mark.asyncio decorator
+        - Tests creating files/directories must have cleanup fixtures
+        - Prefer tmp_path over manual temp directories, and monkeypatch over
+          direct mutation of module state, so tests stay isolated.
+        - Tests must not perform real network I/O; HTTP belongs behind
+          httpx.MockTransport.
   auto_review:
     enabled: true
     auto_incremental_review: true
@@ -89,12 +108,13 @@ reviews:
       - "DO NOT REVIEW"
       - "WIP"
     drafts: false
-  # No source package and no test suite to generate these for.
+  # The web2md scraper (web2md/src/) and its pytest suite (web2md/tests/) are the
+  # only first-party code here.
   finishing_touches:
     docstrings:
-      enabled: false
+      enabled: true
     unit_tests:
-      enabled: false
+      enabled: true
   pre_merge_checks:
     docstrings:
       mode: "off"
diff --git a/.cspell.json b/.cspell.json
index 4d7c709..ab64c1b 100644
--- a/.cspell.json
+++ b/.cspell.json
@@ -28,6 +28,19 @@
     "fdfind",
     "venv",
     "okflintrc",
-    "pycache"
+    "pycache",
+    "httpx",
+    "pythonpath",
+    "testpaths",
+    "minversion",
+    "addopts",
+    "coderabbit",
+    "pycodestyle",
+    "pyflakes",
+    "pydocstyle",
+    "docstring",
+    "docstrings",
+    "conftest",
+    "monkeypatch"
   ]
 }
\ No newline at end of file
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index 5243e5f..363f4c7 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -7,6 +7,7 @@ on:
     paths: &paths
       - "pi/**"
       - "scripts/**"
+      - "web2md/**"
       - "Makefile"
       - "AGENTS.md"
       - "README.md"
@@ -62,6 +63,29 @@ jobs:
       - name: Lint (markdownlint, shellcheck, ruff, hadolint)
         run: make lint MARKDOWNLINT='npx --yes markdownlint-cli2' RUFF='uv run --only-group dev ruff'
 
+  # Unit-test the web2md scraper. Reuses `make test` so the definition stays in
+  # the Makefile. `--only-group` twice installs just pytest and the four scraper
+  # dependencies from the lockfile: no marker-pdf / torch. The suite is offline
+  # (httpx.MockTransport), so this job needs no network beyond the install.
+  test:
+    runs-on: ubuntu-latest
+    timeout-minutes: 10
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+        with:
+          # No git operations after checkout, so don't leave the job token in
+          # .git/config where a later third-party step could read it.
+          persist-credentials: false
+
+      - name: Install uv
+        uses: astral-sh/setup-uv@v9.0.0
+        with:
+          enable-cache: true
+
+      - name: Test (pytest)
+        run: make test PYTEST='uv run --only-group test --only-group web2md pytest'
+
   # Sandbox-only: validate pi/sandbox/spec.yaml against the current Sandbox Kit
   # schema. Static check — no Docker or login; only the CLI download needs the
   # network.
diff --git a/.gitignore b/.gitignore
index 6b2270d..4834521 100644
--- a/.gitignore
+++ b/.gitignore
@@ -6,10 +6,12 @@
 
 # Python
 .venv/
+__pycache__/
+.pytest_cache/
+.ruff_cache/
 
 # web2md HTML fetch cache
 web2md/cache/
-web2md/__pycache__/
 
 # md input folder
 # md/*
diff --git a/AGENTS.md b/AGENTS.md
index f606cb9..0693f3a 100644
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -19,6 +19,14 @@ including the runtime agent configs. `pdf2md/` is the optional upstream step
 that turns a PDF into Markdown with `marker`; it is manual and not wired into
 the `make` pipeline.
 
+`web2md/` is the other upstream step: a deterministic scraper that fetches a
+website into a single file under `md/`, driven by `make scrape`. It is the only
+first-party Python in the repo — module in `web2md/src/`, pytest suite in
+`web2md/tests/`, gitignored HTML cache in `web2md/cache/`. There is no
+`[build-system]`: the module is run by path and pytest imports it via
+`pythonpath` in `pyproject.toml`. Run `make test` after touching either
+directory; the suite is offline and needs no network.
+
 There are **two independent Pi runtimes**, each self-contained and carrying its
 **own copy** of the Pi config (`AGENTS.md`, `settings.json`, `models.json`,
 `skills/`):
@@ -52,9 +60,11 @@ one today, `compile-wiki`. A new task gets a new skill, not more rules in
 
 ```bash
 make lint            # markdownlint + shellcheck + ruff + hadolint (default goal)
+make test            # pytest, the web2md scraper suite (offline)
 make validate        # validate the sandbox kit spec (runs scripts/validate-spec.sh)
 make wiki-sandbox    # compile the OKF wiki via the sandbox runtime (preferred)
 make wiki-container  # compile the OKF wiki via the container runtime
+make scrape          # fetch the website into md/ as one file (web2md)
 make lint-okf        # lint the generated okf/ wiki (okf-lint via pnpm dlx)
 ```
 
diff --git a/Makefile b/Makefile
index 0bfc026..202f30d 100644
--- a/Makefile
+++ b/Makefile
@@ -13,12 +13,16 @@
 #   HADOLINT      hadolint launcher. Local: the brew-installed command. CI: a
 #                 pinned binary downloaded to PATH (see ci.yml), so the default
 #                 works there too. Config lives in .hadolint.yaml.
+#   PYTEST        pytest launcher. Local: `uv run --group test --group web2md`
+#                 (uses the full project venv). CI: the same groups via
+#                 `--only-group`, which drops the project deps — no marker-pdf.
 MARKDOWNLINT ?= markdownlint-cli2
 RUFF ?= uv run ruff
 HADOLINT ?= hadolint
+PYTEST ?= uv run --group test --group web2md pytest
 
 .DEFAULT_GOAL := lint
-.PHONY: lint lint-okf validate wiki-container wiki-sandbox style-guide
+.PHONY: lint lint-okf validate test wiki-container wiki-sandbox scrape
 
 # Lint tracked Markdown, shell, Python, and the container Dockerfile. One
 # Markdown glob per runtime so deleting a runtime is a one-line removal. The
@@ -50,6 +54,12 @@ lint-okf:
 validate:
 	./scripts/validate-spec.sh
 
+# Unit-test the web2md scraper (web2md/tests/). Offline: HTTP is mocked with
+# httpx.MockTransport, so no test opens a socket. Config is in pyproject.toml,
+# which also puts web2md/src/ on the import path.
+test:
+	$(PYTEST)
+
 # Compile the OKF wiki with the containerised Pi runtime (Docker Compose).
 wiki-container:
 	./scripts/compile-wiki-container.sh
@@ -60,4 +70,4 @@ wiki-sandbox:
 
 # Fetch the website into md/ as one file.
 scrape:
-	uv run --group web2md python web2md/web2md.py
+	uv run --group web2md python web2md/src/web2md.py
diff --git a/README.md b/README.md
index 3bc9185..78c7b76 100644
--- a/README.md
+++ b/README.md
@@ -277,6 +277,7 @@ by name so it survives the `okf/*` rule in `.gitignore`.
 
 ```bash
 make lint       # markdownlint, shellcheck, ruff, hadolint
+make test       # pytest, the web2md scraper suite
 make validate   # check pi/sandbox/spec.yaml against the Sandbox Kit schema
 make lint-okf   # lint the generated wiki
 ```
@@ -299,10 +300,18 @@ Once a sandbox exists, this should print `proxy-managed` rather than your key:
 sbx exec pi-kit -- sh -lc 'echo "$OPENROUTER_API_KEY"'
 ```
 
-Python tooling is thin. A `dev` dependency group holds ruff and nothing else,
-and there is no first-party package yet. When real Python code lands, adopt
-`src/md2okf/` and `tests/` with pytest, add a `make test` target and switch on
-the reserved `test` job in CI.
+Python tooling is thin and split across three dependency groups: `dev` (ruff),
+`test` (pytest) and `web2md` (the scraper's four runtime deps). CI installs one
+group per job with `--only-group`, so neither the lint job nor the test job ever
+pulls in the heavy project dependencies (marker-pdf / torch).
+
+The only first-party Python is the web2md scraper, which follows a per-tool
+layout: the module in [web2md/src/](web2md/src/), its pytest suite in
+[web2md/tests/](web2md/tests/). There is no `[build-system]` and nothing is
+installed — `make scrape` runs the module by path, and pytest imports it through
+`pythonpath` in `pyproject.toml`. `make test` runs the suite, and CI runs the
+same command in its `test` job. The suite is offline, so it needs no network and
+never touches `web2md/cache/`.
 
 ## Starting from a PDF
 
diff --git a/pyproject.toml b/pyproject.toml
index 48add70..6baab0f 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -6,8 +6,12 @@ readme = "README.md"
 requires-python = ">=3.12"
 dependencies = ["marker-pdf>=1.10.2", "psutil>=7.2.2"]
 
+# Groups are kept narrow so a CI job installs only what it needs: lint uses
+# --only-group dev, tests use --only-group test --only-group web2md. Neither
+# pulls in the heavy project dependencies (marker-pdf / torch).
 [dependency-groups]
 dev = ["ruff>=0.8"]
+test = ["pytest>=8.4"]
 web2md = [
   "httpx>=0.28.1",
   "beautifulsoup4>=4.15.0",
@@ -15,8 +19,30 @@ web2md = [
   "markdownify>=1.2.3",
 ]
 
-# No [build-system] / src layout yet — there is no first-party Python package in
-# this repo. Adopt src/md2okf/ + tests/ + pytest when real Python code lands
-# (see the README Development section).
+# No [build-system]: web2md is a single module run by path (`make scrape`), not a
+# distributable package. pytest imports it via `pythonpath` below, so there is no
+# install step and no first-party package to build.
+[tool.pytest.ini_options]
+minversion = "8.0"
+addopts = ["-ra", "--strict-markers", "--strict-config"]
+testpaths = ["web2md/tests"]
+pythonpath = ["web2md/src"]
+
 [tool.ruff]
 target-version = "py312"
+line-length = 100
+
+# These rules are what makes the Python review instructions in .coderabbit.yaml
+# enforceable rather than advisory: pycodestyle/pyflakes (E, W, F), import order
+# (I), modern syntax (UP), likely bugs including mutable default arguments (B),
+# needless complexity (SIM, C4, RET), pytest idioms (PT), and docstrings (D).
+[tool.ruff.lint]
+select = ["E", "W", "F", "I", "UP", "B", "SIM", "C4", "RET", "PT", "D"]
+
+[tool.ruff.lint.pydocstyle]
+convention = "google"
+
+[tool.ruff.lint.per-file-ignores]
+# A test's name is its documentation; a docstring on each of them would be noise.
+# Fixtures and helpers are still expected to explain themselves.
+"web2md/tests/*" = ["D100", "D103"]
diff --git a/uv.lock b/uv.lock
index 7de2d79..baf261d 100644
--- a/uv.lock
+++ b/uv.lock
@@ -560,6 +560,15 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/1e/5e/d4e9f1a599fb8e573b7b87160658329fbf28d19eac2718f51fc3def3aa5a/idna-3.18-py3-none-any.whl", hash = "sha256:7f952cbe720b688055e3f87de14f5c3e5fdaa8bc3928985c4077ca689de849a2", size = 65455, upload-time = "2026-06-02T14:34:06.319Z" },
 ]
 
+[[package]]
+name = "iniconfig"
+version = "2.3.0"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/72/34/14ca021ce8e5dfedc35312d08ba8bf51fdd999c576889fc2c24cb97f4f10/iniconfig-2.3.0.tar.gz", hash = "sha256:c76315c77db068650d49c5b56314774a7804df16fee4402c1f19d6d15d8c4730", size = 20503, upload-time = "2025-10-18T21:55:43.219Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/cb/b1/3846dd7f199d53cb17f49cba7e651e9ce294d8497c8c150530ed11865bb8/iniconfig-2.3.0-py3-none-any.whl", hash = "sha256:f631c04d2c48c52b84d0d0549c99ff3859c98df65b3101406327ecc7d53fbf12", size = 7484, upload-time = "2025-10-18T21:55:41.639Z" },
+]
+
 [[package]]
 name = "jinja2"
 version = "3.1.6"
@@ -871,6 +880,9 @@ dependencies = [
 dev = [
     { name = "ruff" },
 ]
+test = [
+    { name = "pytest" },
+]
 web2md = [
     { name = "beautifulsoup4" },
     { name = "httpx" },
@@ -886,6 +898,7 @@ requires-dist = [
 
 [package.metadata.requires-dev]
 dev = [{ name = "ruff", specifier = ">=0.8" }]
+test = [{ name = "pytest", specifier = ">=8.4" }]
 web2md = [
     { name = "beautifulsoup4", specifier = ">=4.15.0" },
     { name = "httpx", specifier = ">=0.28.1" },
@@ -1232,6 +1245,15 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/7d/68/d8d58938dfb1370b266a1a729e6d77a985be23689a0496498ee17b2cbf90/platformdirs-4.11.0-py3-none-any.whl", hash = "sha256:360ccded2b7fce0af0ff80cc8f5942a1c5d99b0e856033acb030bfc634709e74", size = 23247, upload-time = "2026-07-21T13:09:35.422Z" },
 ]
 
+[[package]]
+name = "pluggy"
+version = "1.6.0"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/f9/e2/3e91f31a7d2b083fe6ef3fa267035b518369d9511ffab804f839851d2779/pluggy-1.6.0.tar.gz", hash = "sha256:7dcc130b76258d33b90f61b658791dede3486c3e6bfb003ee5c9bfb396dd22f3", size = 69412, upload-time = "2025-05-15T12:30:07.975Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/54/20/4d324d65cc6d9205fabedc306948156824eb9f0ee1633355a8f7ec5c66bf/pluggy-1.6.0-py3-none-any.whl", hash = "sha256:e920276dd6813095e9377c0bc5566d94c932c33b27a3e3945d8389c374dd4746", size = 20538, upload-time = "2025-05-15T12:30:06.134Z" },
+]
+
 [[package]]
 name = "psutil"
 version = "7.2.2"
@@ -1432,6 +1454,22 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/10/d5/21bac39125df8a93e99c04583486b58a62b5997d6b3541e3ad0f69053392/pypdfium2-5.10.1-py3-none-win_arm64.whl", hash = "sha256:e3301c2f7a66fb8cb57dba857d0c9e90215e178f6602a87c5a306cd98513dab8", size = 3600043, upload-time = "2026-06-15T10:09:14.606Z" },
 ]
 
+[[package]]
+name = "pytest"
+version = "9.1.1"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "colorama", marker = "sys_platform == 'win32'" },
+    { name = "iniconfig" },
+    { name = "packaging" },
+    { name = "pluggy" },
+    { name = "pygments" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/e4/47/b9efed96c114afcfa3c9d3fe98a76a1d14c74a9e266d397cf6eb64be5e01/pytest-9.1.1.tar.gz", hash = "sha256:1088fbde8f2b49d95a549a195707afa7a76a3ce9bcadc26b6d71f0ffda5fe313", size = 1636369, upload-time = "2026-06-19T10:58:32.857Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/24/25/1de2678b631f5a49215c6c96fff41ba892b0a34df68d6d80292b1b48aa7f/pytest-9.1.1-py3-none-any.whl", hash = "sha256:37a86b45efb9a47a61a36449063e8e18d0cab3161329fc099eb21783169c4f0c", size = 386536, upload-time = "2026-06-19T10:58:31.347Z" },
+]
+
 [[package]]
 name = "python-dotenv"
 version = "1.2.2"
diff --git a/web2md/README.md b/web2md/README.md
index 037f0bd..58a6b67 100644
--- a/web2md/README.md
+++ b/web2md/README.md
@@ -9,6 +9,17 @@ guide](https://developers.google.com/style). The output is one file under
 The result is a dated snapshot of a living document. Re-run with `--refresh` to
 update it.
 
+## Layout
+
+| Path | Contents |
+| --- | --- |
+| `src/web2md.py` | the scraper — a single module, run by path, not installed |
+| `tests/` | the pytest suite (see below) |
+| `cache/` | fetched HTML, gitignored; reused unless you pass `--refresh` |
+
+There is no `[build-system]` and no installable package. pytest imports the
+module through `pythonpath = ["web2md/src"]` in the repo's `pyproject.toml`.
+
 ## Fetch and convert
 
 ```bash
@@ -19,10 +30,24 @@ uv sync --group web2md
 make scrape
 
 # Or call the module directly:
-uv run --group web2md python web2md/web2md.py
-uv run --group web2md python web2md/web2md.py --refresh
+uv run --group web2md python web2md/src/web2md.py
+uv run --group web2md python web2md/src/web2md.py --refresh
+```
+
+## Tests
+
+```bash
+make test                                              # the whole suite
+uv run --group test --group web2md pytest web2md/tests # the same, directly
 ```
 
+The suite is fully offline: HTTP is served by `httpx.MockTransport`, so no test
+opens a socket, and the only files written go to pytest's `tmp_path`. It covers
+the pure helpers (slugs, anchors, link rewriting, the Markdown converter,
+assembly, and every `validate_output` error branch), the fetch retry and caching
+logic, and one end-to-end `run()` over a synthetic 72-page site. CI runs it in
+the `test` job.
+
 ## Markdown linting
 
 The generated file under `md/` can be checked manually with the same tools as
diff --git a/web2md/web2md.py b/web2md/src/web2md.py
similarity index 80%
rename from web2md/web2md.py
rename to web2md/src/web2md.py
index d41ca6b..6e202e7 100644
--- a/web2md/web2md.py
+++ b/web2md/src/web2md.py
@@ -11,6 +11,7 @@
 from datetime import UTC, datetime
 from html import unescape
 from pathlib import Path
+from typing import Any
 from urllib.parse import urljoin, urlparse
 
 import httpx
@@ -58,19 +59,28 @@
 SIZE_MIN = 0.4 * 1024 * 1024
 SIZE_MAX = 1.2 * 1024 * 1024
 
-ROOT = Path(__file__).resolve().parent
+# web2md/ — the module itself lives one level down, in web2md/src/.
+ROOT = Path(__file__).resolve().parents[1]
 DEFAULT_CACHE = ROOT / "cache"
 DEFAULT_OUTPUT = ROOT.parent / "md" / "GoogleDeveloperDocumentationStyleGuide.md"
 
 _SANITIZE_RE = re.compile(r"[^A-Za-z0-9._-]")
 _MD_LINK_RE = re.compile(r"\[([^\]]*)\]\((#[^)]+)\)")
-_HEADING_RE = re.compile(r"^(#{1,6})\s+", re.MULTILINE)
+# Unbounded run of '#' on purpose: validate_output filters on len(group) > 6, so
+# capping the group at 6 would make that check unreachable.
+_HEADING_RE = re.compile(r"^(#+)\s+", re.MULTILINE)
 _PANDOC_DD_RE = re.compile(r"^:\s{3,}", re.MULTILINE)
 _DEVSITE_RE = re.compile(r"devsite-|material-icons", re.IGNORECASE)
 
 
 @dataclass(frozen=True)
 class Page:
+    """One page of the book, as listed in the site navigation.
+
+    `slug` is the anchor this page gets in the assembled document; `path` is the
+    site-relative URL path used to resolve cross-page links.
+    """
+
     section: str
     title: str
     slug: str
@@ -79,11 +89,18 @@ class Page:
 
 
 def slug_from_path(path: str) -> str:
+    """Derive a page slug from its URL path; the book root becomes "style"."""
     rest = path.removeprefix("/style").strip("/")
     return rest or "style"
 
 
 def sanitize_id(raw: str) -> str:
+    """Reduce an HTML id to `[A-Za-z0-9._-]`, escaping anything else as uXXXX.
+
+    The escape is reversible enough to stay unique, so two distinct ids never
+    collapse onto the same anchor.
+    """
+
     def repl(match: re.Match[str]) -> str:
         return f"u{ord(match.group(0)):04x}"
 
@@ -92,16 +109,23 @@ def repl(match: re.Match[str]) -> str:
 
 
 def namespaced_anchor(page_slug: str, raw_id: str) -> str:
+    """Scope an id to its page, so ids repeated across pages stay distinct."""
     return f"{page_slug}--{sanitize_id(raw_id)}"
 
 
 def nav_text(el: Tag) -> str:
+    """Read a nav entry's label, preferring its `.devsite-nav-text` span."""
     span = el.select_one(".devsite-nav-text")
     text = span.get_text(" ", strip=True) if span else el.get_text(" ", strip=True)
     return unescape(text)
 
 
 def discover_pages(html: str) -> list[Page]:
+    """List every book page from the nav, in document order.
+
+    Raises SystemExit if the nav is missing or its shape looks wrong, so a
+    silently truncated book fails the run instead of producing a short document.
+    """
     soup = BeautifulSoup(html, "lxml")
     nav = soup.select_one(NAV)
     if nav is None:
@@ -160,6 +184,13 @@ def fetch_html(
     *,
     refresh: bool,
 ) -> str:
+    """Return the page HTML, from `cache_path` when possible.
+
+    On a cache miss the request is retried with exponential backoff, bounded by
+    MAX_RETRIES and capped at 30s per wait, honouring `Retry-After` on HTTP 429.
+    Only a successful response is cached. Raises SystemExit once the retry
+    budget is spent.
+    """
     if cache_path.exists() and not refresh:
         return cache_path.read_text(encoding="utf-8")
 
@@ -198,8 +229,11 @@ def fetch_html(
 
 
 def replace_icon_spans(body: Tag) -> None:
-    # Only rewrite icons inside definition terms. The word-list legend uses the
-    # same empty spans next to prose that already explains them.
+    """Turn DevSite's empty icon spans into the words they stand for.
+
+    Only inside definition terms: the word-list legend uses the same spans next
+    to prose that already explains them, where the replacement would duplicate.
+    """
     for class_name, label in ICON_TEXT.items():
         for el in body.select(f".{class_name}"):
             if el.find_parent("dt") is None:
@@ -209,6 +243,7 @@ def replace_icon_spans(body: Tag) -> None:
 
 
 def unwrap_devsite_code(body: Tag) -> None:
+    """Replace each <devsite-code> wrapper with the <pre> it holds."""
     for el in body.find_all("devsite-code"):
         pre = el.find("pre")
         if pre is not None:
@@ -218,6 +253,7 @@ def unwrap_devsite_code(body: Tag) -> None:
 
 
 def drop_noise(body: Tag, dropped_classes: set[str]) -> None:
+    """Strip site chrome listed in DROP, recording removed classes for the log."""
     for selector in DROP:
         for el in body.select(selector):
             for cls in el.get("class") or []:
@@ -231,6 +267,11 @@ def drop_noise(body: Tag, dropped_classes: set[str]) -> None:
 
 
 def clean_body(html: str, page: Page, dropped_classes: set[str]) -> Tag:
+    """Extract the article body and strip it down to publishable content.
+
+    Raises SystemExit if the body is missing, or if its headings fall outside
+    the h2–h4 range the assembled document's +2 shift assumes.
+    """
     soup = BeautifulSoup(html, "lxml")
     body = soup.select_one(BODY)
     if body is None:
@@ -240,19 +281,21 @@ def clean_body(html: str, page: Page, dropped_classes: set[str]) -> Tag:
     unwrap_devsite_code(body)
     drop_noise(body, dropped_classes)
 
-    # Heading depth check before conversion (body uses h2–h4).
+    # Heading depth check before conversion. The body is expected to use h2–h4:
+    # deeper would overflow h6 once assemble() shifts everything down by two,
+    # and an h1 would compete with the document title.
     for heading in body.find_all(re.compile(r"^h[1-6]$")):
         level = int(heading.name[1])
         if level > 4:
             raise SystemExit(f"body heading deeper than h4 on {page.url}: <{heading.name}>")
         if level < 2 and heading.get("id") != "key-takeaways-panel-title":
-            # h1 should not appear in the article body for this book.
-            pass
+            raise SystemExit(f"unexpected <h1> in the article body on {page.url}")
 
     return body
 
 
 def collect_ids(body: Tag, page_slug: str, anchor_map: dict[tuple[str, str], str]) -> None:
+    """Record every id on the page as (page_slug, raw_id) -> namespaced anchor."""
     for el in body.find_all(attrs={"id": True}):
         raw_id = el["id"]
         if not raw_id or raw_id == "key-takeaways-panel-title":
@@ -262,6 +305,7 @@ def collect_ids(body: Tag, page_slug: str, anchor_map: dict[tuple[str, str], str
 
 
 def absolutize_url(href: str) -> str:
+    """Resolve a possibly site-relative href against the site base URL."""
     return urljoin(BASE, href)
 
 
@@ -272,6 +316,12 @@ def rewrite_internal_href(
     anchor_map: dict[tuple[str, str], str],
     unresolved: list[str],
 ) -> str:
+    """Rewrite one href for the single-document output.
+
+    In-book targets become `#anchor` fragments; anything else is left alone or
+    made absolute. Targets that cannot be resolved are appended to `unresolved`
+    and fall back to an absolute URL, so no link is silently dropped.
+    """
     parsed = urlparse(href)
     fragment = parsed.fragment
 
@@ -325,6 +375,11 @@ def apply_anchors_and_links(
     anchor_map: dict[tuple[str, str], str],
     unresolved: list[str],
 ) -> None:
+    """Emit explicit anchors for id-bearing elements and rewrite links in place.
+
+    Mutates `body`: ids move onto injected <a id="…"> tags, hrefs are rewritten
+    for the single-document output, and image sources are made absolute.
+    """
     # Inject an explicit anchor before every id-bearing element (sections,
     # headings, dt terms, …). markdownify drops ids on non-heading tags, and
     # many DevSite fragments live on <section id="…"> wrappers. DevSite often
@@ -358,17 +413,21 @@ class StyleGuideConverter(MarkdownConverter):
     """DevSite-specific HTML → Markdown rules."""
 
     class Options(MarkdownConverter.Options):
+        """markdownify settings this book is converted with."""
+
         heading_style = ATX
         bullets = "-"
         escape_asterisks = False
         escape_underscores = False
 
-    def __init__(self, page_slug: str, aside_warnings: list[str], **options):
+    def __init__(self, page_slug: str, aside_warnings: list[str], **options: Any) -> None:
+        """Bind the converter to one page; `aside_warnings` collects warnings."""
         super().__init__(**options)
         self.page_slug = page_slug
         self.aside_warnings = aside_warnings
 
     def convert_aside(self, el: Tag, text: str, parent_tags: set[str]) -> str:
+        """Render an <aside> as a GitHub callout, defaulting to NOTE."""
         classes = el.get("class") or []
         kind = "NOTE"
         matched = False
@@ -386,22 +445,35 @@ def convert_aside(self, el: Tag, text: str, parent_tags: set[str]) -> str:
         return f"\n\n> [!{kind}]\n> {quoted}\n\n"
 
     def convert_a(self, el: Tag, text: str, parent_tags: set[str]) -> str:
-        # Keep explicit fragment targets; markdownify would otherwise drop empty
-        # anchors that have an id and no href.
+        """Convert a link, preserving the bare <a id="…"> anchors we injected.
+
+        markdownify would otherwise drop an empty anchor that has an id and no
+        href, taking every cross-reference target with it.
+        """
         anchor_id = el.get("id")
         if anchor_id and not el.get("href") and not (text or "").strip():
             return f'<a id="{anchor_id}"></a>'
         return super().convert_a(el, text, parent_tags)
 
-    def process_text(self, el, parent_tags=None):
+    def process_text(
+        self, el: NavigableString, parent_tags: set[str] | None = None
+    ) -> str:
+        """Escape literal Markdown fences that appear in prose.
+
+        DevSite documents mention ``` as prose; left alone those open a fence
+        and swallow the content that follows.
+        """
         text = super().process_text(el, parent_tags=parent_tags)
-        # DevSite docs mention Markdown fences as literal ``` in prose; left
-        # alone those open a fence and swallow the following content.
         if parent_tags is None or "pre" not in parent_tags:
             text = text.replace("```", r"\`\`\`")
         return text
 
     def convert_dt(self, el: Tag, text: str, parent_tags: set[str]) -> str:
+        """Render a definition term as a bold line.
+
+        Markdown has no definition lists; bold term plus indented body keeps the
+        structure readable and avoids Pandoc-only syntax.
+        """
         text = re.sub(r"\s+", " ", (text or "").strip())
         if "_inline" in parent_tags:
             return f" {text} "
@@ -410,6 +482,7 @@ def convert_dt(self, el: Tag, text: str, parent_tags: set[str]) -> str:
         return f"\n\n**{text}**\n"
 
     def convert_dd(self, el: Tag, text: str, parent_tags: set[str]) -> str:
+        """Render a definition body as a four-space indented block."""
         text = (text or "").strip()
         if "_inline" in parent_tags:
             return f" {text} "
@@ -421,9 +494,12 @@ def convert_dd(self, el: Tag, text: str, parent_tags: set[str]) -> str:
         return f"\n{indented}\n"
 
     def convert_hN(self, n: int, el: Tag, text: str, parent_tags: set[str]) -> str:
+        """Shift a body heading down two levels, clamped to h6.
+
+        The document already spends # / ## / ### on title / section / page.
+        """
         if "_inline" in parent_tags:
             return text
-        # Document uses # / ## / ### for title / section / page, so body shifts +2.
         n = max(1, min(6, n + 2))
         text = re.sub(r"\s+", " ", (text or "").strip())
         if not text:
@@ -431,6 +507,7 @@ def convert_hN(self, n: int, el: Tag, text: str, parent_tags: set[str]) -> str:
         return f"\n\n{'#' * n} {text}\n\n"
 
     def convert_img(self, el: Tag, text: str, parent_tags: set[str]) -> str:
+        """Render an image, degrading to its alt text when there is no source."""
         alt = el.get("alt") or ""
         src = el.get("src") or ""
         if not src:
@@ -438,6 +515,7 @@ def convert_img(self, el: Tag, text: str, parent_tags: set[str]) -> str:
         return f"![{alt}]({src})"
 
     def convert_pre(self, el: Tag, text: str, parent_tags: set[str]) -> str:
+        """Render a code block as an unlabelled fence, always newline-terminated."""
         code = el.get_text().removeprefix("\n")
         if not code.endswith("\n"):
             code = code + "\n"
@@ -445,11 +523,13 @@ def convert_pre(self, el: Tag, text: str, parent_tags: set[str]) -> str:
 
 
 def convert_body(body: Tag, page_slug: str, aside_warnings: list[str]) -> str:
+    """Convert one cleaned article body to Markdown."""
     converter = StyleGuideConverter(page_slug=page_slug, aside_warnings=aside_warnings)
     return converter.convert(str(body)).strip()
 
 
 def build_toc(pages: list[Page]) -> str:
+    """Render the table of contents, grouped by section in nav order."""
     lines = ["## Table of contents", ""]
     current_section = None
     for page in pages:
@@ -465,6 +545,7 @@ def build_toc(pages: list[Page]) -> str:
 
 
 def assemble(pages: list[Page], bodies: dict[str, str], timestamp: str) -> str:
+    """Join frontmatter, table of contents, and every page body into one file."""
     parts: list[str] = [
         "---",
         "type: Website",
@@ -512,6 +593,11 @@ def markdown_without_fences(md: str) -> str:
 
 
 def validate_output(md: str, pages: list[Page], bodies: dict[str, str]) -> None:
+    """Check the assembled document before it is written.
+
+    Every failure is collected and reported together, then raises SystemExit, so
+    one run surfaces all the problems rather than only the first.
+    """
     errors: list[str] = []
 
     if len(bodies) != len(pages):
@@ -576,6 +662,11 @@ def validate_output(md: str, pages: list[Page], bodies: dict[str, str]) -> None:
 
 
 def run(*, refresh: bool, cache_dir: Path, output: Path) -> None:
+    """Fetch, clean, convert, validate, and write the whole book.
+
+    Nothing is written to `output` unless validation passes, so a failed run
+    leaves the previous snapshot intact.
+    """
     warnings.filterwarnings("ignore", category=UserWarning, module="bs4")
     cache_dir.mkdir(parents=True, exist_ok=True)
     output.parent.mkdir(parents=True, exist_ok=True)
@@ -641,6 +732,7 @@ def run(*, refresh: bool, cache_dir: Path, output: Path) -> None:
 
 
 def main(argv: list[str] | None = None) -> None:
+    """Parse the command line and run the scrape."""
     parser = argparse.ArgumentParser(
         description="Fetch developers.google.com/style into one Markdown file."
     )
diff --git a/web2md/tests/conftest.py b/web2md/tests/conftest.py
new file mode 100644
index 0000000..72c554c
--- /dev/null
+++ b/web2md/tests/conftest.py
@@ -0,0 +1,98 @@
+"""Shared fixtures for the web2md tests.
+
+Every test in this suite is offline: nothing here or in the test modules opens a
+socket. HTTP is exercised through `httpx.MockTransport`, and the only filesystem
+writes go to pytest's `tmp_path`.
+"""
+
+from __future__ import annotations
+
+from collections.abc import Callable
+
+import pytest
+from bs4 import BeautifulSoup, Tag
+
+import web2md
+
+
+@pytest.fixture(autouse=True)
+def no_sleep(monkeypatch: pytest.MonkeyPatch) -> None:
+    """Make the retry backoff instant so fetch tests don't idle for seconds."""
+    monkeypatch.setattr(web2md.time, "sleep", lambda _seconds: None)
+
+
+@pytest.fixture
+def frag() -> Callable[[str], Tag]:
+    """Parse an HTML fragment into a <div> Tag that is still attached to a soup.
+
+    Attachment matters: `apply_anchors_and_links` walks up to the BeautifulSoup
+    root to call `new_tag`, which a detached Tag does not provide.
+    """
+
+    def _frag(html: str) -> Tag:
+        return BeautifulSoup(f"<div>{html}</div>", "lxml").div
+
+    return _frag
+
+
+@pytest.fixture
+def to_md(frag: Callable[[str], Tag]) -> Callable[..., str]:
+    """Convert an HTML fragment through the real converter entry point."""
+
+    def _to_md(html: str, *, slug: str = "page", warnings: list[str] | None = None) -> str:
+        return web2md.convert_body(frag(html), slug, warnings if warnings is not None else [])
+
+    return _to_md
+
+
+@pytest.fixture
+def make_page() -> Callable[..., web2md.Page]:
+    def _make_page(
+        slug: str = "word-list",
+        *,
+        section: str = "Reference",
+        title: str = "Word list",
+        path: str | None = None,
+    ) -> web2md.Page:
+        if path is None:
+            path = "/style" if slug == "style" else f"/style/{slug}"
+        return web2md.Page(
+            section=section,
+            title=title,
+            slug=slug,
+            url=f"{web2md.BASE}{path}",
+            path=path,
+        )
+
+    return _make_page
+
+
+def build_nav_html(sections: int = 6, per_section: int = 12, extra_items: str = "") -> str:
+    """Build a DevSite book nav big enough to clear `discover_pages`' guards.
+
+    The leading item sits above the first heading, standing in for the
+    book-picker tabs that `discover_pages` is meant to skip.
+    """
+    parts = [
+        f'<div class="{web2md.NAV.lstrip(".")}"><ul>',
+        '<li class="devsite-nav-item"><a href="/docs/guides">Guides</a></li>',
+    ]
+    for section in range(sections):
+        parts.append(
+            '<li class="devsite-nav-heading">'
+            f'<span class="devsite-nav-text">Section {section}</span></li>'
+        )
+        for item in range(per_section):
+            parts.append(
+                '<li class="devsite-nav-item">'
+                f'<a href="/style/page-{section}-{item}">'
+                f'<span class="devsite-nav-text">Page {section}.{item}</span></a></li>'
+            )
+    parts.append(extra_items)
+    parts.append("</ul></div>")
+    return "".join(parts)
+
+
+@pytest.fixture
+def nav_html() -> Callable[..., str]:
+    return build_nav_html
diff --git a/web2md/tests/test_assemble.py b/web2md/tests/test_assemble.py
new file mode 100644
index 0000000..4aee898
--- /dev/null
+++ b/web2md/tests/test_assemble.py
@@ -0,0 +1,102 @@
+"""Table of contents, document assembly, and the fence-stripping helper."""
+
+from __future__ import annotations
+
+from collections.abc import Callable
+
+import web2md
+
+TIMESTAMP = "2026-08-01T12:00:00Z"
+
+
+def _book(make_page: Callable[..., web2md.Page]) -> list[web2md.Page]:
+    return [
+        make_page("style", section="Overview", title="About", path="/style"),
+        make_page("tense", section="Language", title="Tense"),
+        make_page("voice", section="Language", title="Voice"),
+    ]
+
+
+def test_build_toc_groups_pages_under_their_section(
+    make_page: Callable[..., web2md.Page],
+) -> None:
+    toc = web2md.build_toc(_book(make_page))
+
+    assert toc.splitlines() == [
+        "## Table of contents",
+        "",
+        "### Overview",
+        "",
+        "- [About](#style)",
+        "",
+        "### Language",
+        "",
+        "- [Tense](#tense)",
+        "- [Voice](#voice)",
+    ]
+    assert toc.endswith("\n")
+
+
+def test_assemble_writes_frontmatter_and_timestamp(
+    make_page: Callable[..., web2md.Page],
+) -> None:
+    pages = _book(make_page)
+    bodies = {p.slug: f"Body of {p.slug}." for p in pages}
+
+    md = web2md.assemble(pages, bodies, timestamp=TIMESTAMP)
+
+    lines = md.splitlines()
+    assert lines[0] == "---"
+    assert "type: Website" in lines
+    assert f"timestamp: {TIMESTAMP}" in lines
+    assert f"resource: {web2md.BOOK}" in lines
+    assert "# Google Developer Documentation Style Guide" in lines
+    assert f"*Snapshot of [{web2md.BOOK}]({web2md.BOOK}) generated 2026-08-01.*" in lines
+
+
+def test_assemble_emits_each_section_once_with_page_anchors(
+    make_page: Callable[..., web2md.Page],
+) -> None:
+    pages = _book(make_page)
+    bodies = {p.slug: f"Body of {p.slug}." for p in pages}
+
+    md = web2md.assemble(pages, bodies, timestamp=TIMESTAMP)
+
+    # Anchored so the TOC's "### Language" heading is not counted too.
+    assert md.count("\n## Language\n") == 1
+    for page in pages:
+        assert f'<a id="{page.slug}"></a>' in md
+        assert f"### {page.title}" in md
+        assert f"*Source: <{page.url}>*" in md
+        assert bodies[page.slug] in md
+
+
+def test_assemble_drops_the_trailing_divider(make_page: Callable[..., web2md.Page]) -> None:
+    pages = _book(make_page)
+    bodies = {p.slug: f"Body of {p.slug}." for p in pages}
+
+    md = web2md.assemble(pages, bodies, timestamp=TIMESTAMP)
+
+    assert md.endswith("Body of voice.\n")
+    assert not md.rstrip().endswith("---")
+    # Two dividers between three pages, plus the frontmatter's closing fence.
+    # The opening fence is at offset 0, so it has no leading newline to match.
+    assert md.count("\n---\n") == 3
+
+
+def test_markdown_without_fences_strips_only_fenced_blocks() -> None:
+    md = "before\n```\ndevsite-toc\n```\nafter"
+
+    assert web2md.markdown_without_fences(md) == "before\n\nafter"
+
+
+def test_markdown_without_fences_handles_several_blocks() -> None:
+    md = "a\n```\none\n```\nb\n```\ntwo\n```\nc"
+
+    stripped = web2md.markdown_without_fences(md)
+
+    assert "one" not in stripped
+    assert "two" not in stripped
+    assert "a" in stripped
+    assert "b" in stripped
+    assert "c" in stripped
diff --git a/web2md/tests/test_cleaning.py b/web2md/tests/test_cleaning.py
new file mode 100644
index 0000000..253b406
--- /dev/null
+++ b/web2md/tests/test_cleaning.py
@@ -0,0 +1,182 @@
+"""DOM cleaning: icon spans, devsite-code, noise removal, body extraction, ids."""
+
+from __future__ import annotations
+
+from collections.abc import Callable
+
+import pytest
+from bs4 import Tag
+
+import web2md
+
+
+def _body(inner: str) -> str:
+    return f'<html><body><div class="devsite-article-body">{inner}</div></body></html>'
+
+
+def test_icon_spans_inside_dt_become_text(frag: Callable[[str], Tag]) -> None:
+    body = frag(
+        '<dl><dt><span class="icon-dontuse"></span>whitelist</dt>'
+        "<dd>Use allowlist.</dd></dl>"
+    )
+    web2md.replace_icon_spans(body)
+
+    assert "Don't use: whitelist" in body.get_text(" ", strip=True)
+
+
+def test_icon_spans_outside_dt_are_left_alone(frag: Callable[[str], Tag]) -> None:
+    # The word-list legend uses the same empty spans next to prose that already
+    # explains them; rewriting there would duplicate the explanation.
+    body = frag('<p><span class="icon-dontuse"></span> means don\'t use this term.</p>')
+    web2md.replace_icon_spans(body)
+
+    assert "Don't use:" not in body.get_text(" ", strip=True)
+
+
+@pytest.mark.parametrize(
+    ("class_name", "label"),
+    sorted(web2md.ICON_TEXT.items()),
+)
+def test_every_icon_mapping_is_applied(
+    frag: Callable[[str], Tag], class_name: str, label: str
+) -> None:
+    body = frag(f'<dl><dt><span class="{class_name}"></span>term</dt></dl>')
+    web2md.replace_icon_spans(body)
+
+    assert body.dt.get_text(" ", strip=True).startswith(label)
+
+
+def test_devsite_code_is_replaced_by_its_pre(frag: Callable[[str], Tag]) -> None:
+    body = frag("<devsite-code><pre>print(1)</pre></devsite-code>")
+    web2md.unwrap_devsite_code(body)
+
+    assert body.find("devsite-code") is None
+    assert body.pre is not None
+    assert body.pre.get_text() == "print(1)"
+
+
+def test_devsite_code_without_pre_is_unwrapped(frag: Callable[[str], Tag]) -> None:
+    body = frag("<devsite-code>bare text</devsite-code>")
+    web2md.unwrap_devsite_code(body)
+
+    assert body.find("devsite-code") is None
+    assert body.get_text() == "bare text"
+
+
+@pytest.mark.parametrize("selector", web2md.DROP)
+def test_every_drop_selector_removes_its_element(
+    frag: Callable[[str], Tag], selector: str
+) -> None:
+    if selector.startswith("."):
+        markup = f'<div class="{selector[1:]}">noise</div>'
+    elif selector.startswith("["):
+        markup = '<div aria-hidden="true">noise</div>'
+    else:
+        markup = f"<{selector}>noise</{selector}>"
+    body = frag(f"<p>keep</p>{markup}")
+
+    web2md.drop_noise(body, set())
+
+    assert "noise" not in body.get_text()
+    assert "keep" in body.get_text()
+
+
+def test_drop_noise_records_the_class_values_it_removed(frag: Callable[[str], Tag]) -> None:
+    body = frag('<div class="devsite-rating-container extra">noise</div>')
+    dropped: set[str] = set()
+
+    web2md.drop_noise(body, dropped)
+
+    assert dropped == {"devsite-rating-container", "extra"}
+
+
+def test_empty_permalink_anchors_are_dropped_but_image_links_survive(
+    frag: Callable[[str], Tag],
+) -> None:
+    body = frag(
+        '<a href="#section"></a>'
+        '<a href="#other"><img src="/i.png" alt="pic"></a>'
+        '<a href="/style/tense">Tense</a>'
+    )
+    web2md.drop_noise(body, set())
+
+    hrefs = [a["href"] for a in body.find_all("a", href=True)]
+    assert hrefs == ["#other", "/style/tense"]
+
+
+def test_clean_body_runs_the_full_pipeline(make_page: Callable[..., web2md.Page]) -> None:
+    html = _body(
+        '<h2 id="intro">Intro</h2>'
+        '<devsite-code><pre>code</pre></devsite-code>'
+        '<devsite-feedback>rate me</devsite-feedback>'
+        '<dl><dt><span class="icon-avoid"></span>term</dt></dl>'
+    )
+    dropped: set[str] = set()
+
+    body = web2md.clean_body(html, make_page(), dropped)
+
+    text = body.get_text(" ", strip=True)
+    assert "rate me" not in text
+    assert "Avoid: term" in text
+    assert body.find("devsite-code") is None
+
+
+def test_clean_body_without_article_body_is_fatal(make_page: Callable[..., web2md.Page]) -> None:
+    with pytest.raises(SystemExit, match="missing article body"):
+        web2md.clean_body("<html><body><p>nope</p></body></html>", make_page(), set())
+
+
+@pytest.mark.parametrize("tag", ["h5", "h6"])
+def test_clean_body_rejects_headings_deeper_than_h4(
+    make_page: Callable[..., web2md.Page], tag: str
+) -> None:
+    with pytest.raises(SystemExit, match="heading deeper than h4"):
+        web2md.clean_body(_body(f"<{tag}>too deep</{tag}>"), make_page(), set())
+
+
+def test_clean_body_rejects_an_h1_that_would_rival_the_document_title(
+    make_page: Callable[..., web2md.Page],
+) -> None:
+    with pytest.raises(SystemExit, match="unexpected <h1>"):
+        web2md.clean_body(_body("<h1>Page title</h1>"), make_page(), set())
+
+
+def test_clean_body_allows_the_key_takeaways_panel_title(
+    make_page: Callable[..., web2md.Page],
+) -> None:
+    # DevSite gives that panel an h1; it is dropped as noise, not rejected.
+    html = _body('<h1 id="key-takeaways-panel-title">Key takeaways</h1><h2>Real</h2>')
+
+    body = web2md.clean_body(html, make_page(), set())
+
+    assert "Real" in body.get_text()
+
+
+@pytest.mark.parametrize("tag", ["h2", "h3", "h4"])
+def test_clean_body_accepts_the_expected_heading_range(
+    make_page: Callable[..., web2md.Page], tag: str
+) -> None:
+    body = web2md.clean_body(_body(f"<{tag}>Fine</{tag}>"), make_page(), set())
+
+    assert body.find(tag) is not None
+
+
+def test_collect_ids_namespaces_every_id(frag: Callable[[str], Tag]) -> None:
+    body = frag('<section id="terms"><h2 id="terms">Terms</h2><dt id="a b">x</dt></section>')
+    anchor_map: dict[tuple[str, str], str] = {}
+
+    web2md.collect_ids(body, "word-list", anchor_map)
+
+    assert anchor_map == {
+        ("word-list", "terms"): "word-list--terms",
+        ("word-list", "a b"): "word-list--au0020b",
+    }
+
+
+def test_collect_ids_skips_the_key_takeaways_panel(frag: Callable[[str], Tag]) -> None:
+    body = frag('<h2 id="key-takeaways-panel-title">Key takeaways</h2>')
+    anchor_map: dict[tuple[str, str], str] = {}
+
+    web2md.collect_ids(body, "page", anchor_map)
+
+    assert anchor_map == {}
diff --git a/web2md/tests/test_cli.py b/web2md/tests/test_cli.py
new file mode 100644
index 0000000..9f5e00e
--- /dev/null
+++ b/web2md/tests/test_cli.py
@@ -0,0 +1,65 @@
+"""Command-line parsing and the default cache / output locations."""
+
+from __future__ import annotations
+
+from pathlib import Path
+
+import pytest
+
+import web2md
+
+# <repo>/web2md/src/web2md.py -> <repo>
+REPO_ROOT = Path(web2md.__file__).resolve().parents[2]
+
+
+def test_default_cache_sits_beside_the_src_directory() -> None:
+    # Guards the move into web2md/src/: a ROOT off by one level would silently
+    # relocate the cache to web2md/src/cache/.
+    assert web2md.DEFAULT_CACHE == REPO_ROOT / "web2md" / "cache"
+
+
+def test_default_output_goes_to_the_repo_level_md_directory() -> None:
+    assert web2md.DEFAULT_OUTPUT == (
+        REPO_ROOT / "md" / "GoogleDeveloperDocumentationStyleGuide.md"
+    )
+
+
+@pytest.fixture
+def calls(monkeypatch: pytest.MonkeyPatch) -> list[dict[str, object]]:
+    recorded: list[dict[str, object]] = []
+    monkeypatch.setattr(web2md, "run", lambda **kwargs: recorded.append(kwargs))
+    return recorded
+
+
+def test_no_arguments_uses_the_defaults(calls: list[dict[str, object]]) -> None:
+    web2md.main([])
+
+    assert calls == [
+        {
+            "refresh": False,
+            "cache_dir": web2md.DEFAULT_CACHE,
+            "output": web2md.DEFAULT_OUTPUT,
+        }
+    ]
+
+
+def test_refresh_flag_is_forwarded(calls: list[dict[str, object]]) -> None:
+    web2md.main(["--refresh"])
+
+    assert calls[0]["refresh"] is True
+
+
+def test_cache_dir_and_output_are_parsed_as_paths(
+    calls: list[dict[str, object]], tmp_path: Path
+) -> None:
+    web2md.main(["--cache-dir", str(tmp_path / "c"), "--output", str(tmp_path / "o.md")])
+
+    assert calls[0]["cache_dir"] == tmp_path / "c"
+    assert calls[0]["output"] == tmp_path / "o.md"
+
+
+def test_unknown_option_exits(calls: list[dict[str, object]]) -> None:
+    with pytest.raises(SystemExit):
+        web2md.main(["--nope"])
+
+    assert calls == []
diff --git a/web2md/tests/test_converter.py b/web2md/tests/test_converter.py
new file mode 100644
index 0000000..87640fe
--- /dev/null
+++ b/web2md/tests/test_converter.py
@@ -0,0 +1,123 @@
+"""`StyleGuideConverter` — the DevSite-specific HTML → Markdown rules."""
+
+from __future__ import annotations
+
+from collections.abc import Callable
+
+import pytest
+
+import web2md
+
+
+@pytest.mark.parametrize(
+    ("css_class", "kind"),
+    sorted(web2md.ASIDE_KIND.items()),
+)
+def test_asides_become_github_callouts(
+    to_md: Callable[..., str], css_class: str, kind: str
+) -> None:
+    result = to_md(f'<aside class="{css_class}"><p>Careful.</p></aside>')
+
+    assert result == f"> [!{kind}]\n> Careful."
+
+
+def test_multi_paragraph_aside_quotes_every_line(to_md: Callable[..., str]) -> None:
+    result = to_md('<aside class="note"><p>Line one.</p><p>Line two.</p></aside>')
+
+    assert result == "> [!NOTE]\n> Line one.\n> Line two."
+
+
+def test_unknown_aside_class_warns_and_defaults_to_note(to_md: Callable[..., str]) -> None:
+    warnings: list[str] = []
+
+    result = to_md('<aside class="mystery"><p>Hm.</p></aside>', slug="tense", warnings=warnings)
+
+    assert result == "> [!NOTE]\n> Hm."
+    assert len(warnings) == 1
+    assert "mystery" in warnings[0]
+    assert "tense" in warnings[0]
+
+
+def test_classless_aside_does_not_warn(to_md: Callable[..., str]) -> None:
+    warnings: list[str] = []
+
+    assert to_md("<aside><p>Hm.</p></aside>", warnings=warnings) == "> [!NOTE]\n> Hm."
+    assert warnings == []
+
+
+def test_empty_aside_produces_nothing(to_md: Callable[..., str]) -> None:
+    assert to_md('<aside class="note"></aside>') == ""
+
+
+def test_explicit_anchor_survives_conversion(to_md: Callable[..., str]) -> None:
+    assert to_md('<a id="word-list--term"></a><p>text</p>').startswith(
+        '<a id="word-list--term"></a>'
+    )
+
+
+def test_ordinary_links_still_convert(to_md: Callable[..., str]) -> None:
+    assert to_md('<p><a href="#tense">Tense</a></p>') == "[Tense](#tense)"
+
+
+def test_literal_fences_in_prose_are_escaped(to_md: Callable[..., str]) -> None:
+    # Left alone, these would open a fence and swallow the following content.
+    assert to_md("<p>use ``` fences</p>") == r"use \`\`\` fences"
+
+
+def test_fences_inside_pre_are_not_escaped(to_md: Callable[..., str]) -> None:
+    assert "\\`" not in to_md("<pre>```</pre>")
+
+
+def test_definition_lists_become_bold_terms_with_indented_bodies(
+    to_md: Callable[..., str],
+) -> None:
+    result = to_md("<dl><dt>term</dt><dd><p>First.</p><p>Second.</p></dd></dl>")
+
+    assert result == "**term**\n    First.\n\n    Second."
+
+
+def test_definition_term_whitespace_is_collapsed(to_md: Callable[..., str]) -> None:
+    assert to_md("<dl><dt>two\n   words</dt></dl>") == "**two words**"
+
+
+def test_empty_definition_parts_collapse(to_md: Callable[..., str]) -> None:
+    assert to_md("<dl><dt></dt><dd></dd></dl>") == ""
+
+
+def test_body_headings_shift_down_by_two(to_md: Callable[..., str]) -> None:
+    # The document already uses # / ## / ### for title / section / page.
+    result = to_md("<h2>Two</h2><p>x</p><h4>Four</h4>")
+
+    assert result == "#### Two\n\nx\n\n###### Four"
+
+
+def test_heading_depth_is_clamped_at_six(to_md: Callable[..., str]) -> None:
+    assert to_md("<h5>Five</h5>") == "###### Five"
+    assert to_md("<h6>Six</h6>") == "###### Six"
+
+
+def test_empty_heading_produces_nothing(to_md: Callable[..., str]) -> None:
+    assert to_md("<h2>   </h2>") == ""
+
+
+def test_images_keep_alt_and_src(to_md: Callable[..., str]) -> None:
+    assert to_md('<img src="https://x/i.png" alt="A chart">') == "![A chart](https://x/i.png)"
+
+
+def test_image_without_src_degrades_to_its_alt_text(to_md: Callable[..., str]) -> None:
+    assert to_md('<img alt="A chart">') == "A chart"
+
+
+def test_pre_becomes_a_fenced_block_with_a_trailing_newline(to_md: Callable[..., str]) -> None:
+    assert to_md("<pre>a = 1</pre>") == "```\na = 1\n```"
+
+
+def test_pre_leading_newline_is_dropped(to_md: Callable[..., str]) -> None:
+    assert to_md("<pre>\na = 1\n</pre>") == "```\na = 1\n```"
+
+
+def test_bullets_use_dashes_and_asterisks_are_not_escaped(to_md: Callable[..., str]) -> None:
+    result = to_md("<ul><li>one</li><li>two</li></ul>")
+
+    assert result == "- one\n- two"
+    assert to_md("<p>a * b _ c</p>") == "a * b _ c"
diff --git a/web2md/tests/test_discovery.py b/web2md/tests/test_discovery.py
new file mode 100644
index 0000000..56ba664
--- /dev/null
+++ b/web2md/tests/test_discovery.py
@@ -0,0 +1,63 @@
+"""`discover_pages` — nav parsing and its guard rails."""
+
+from __future__ import annotations
+
+from collections.abc import Callable
+
+import pytest
+
+import web2md
+
+
+def test_discovers_every_nav_item_with_its_section(nav_html: Callable[..., str]) -> None:
+    pages = web2md.discover_pages(nav_html(sections=6, per_section=12))
+
+    assert len(pages) == 72
+    assert pages[0].section == "Section 0"
+    assert pages[0].title == "Page 0.0"
+    assert pages[0].slug == "page-0-0"
+    assert pages[0].path == "/style/page-0-0"
+    assert pages[0].url == "https://developers.google.com/style/page-0-0"
+    assert {p.section for p in pages} == {f"Section {i}" for i in range(6)}
+
+
+def test_skips_book_picker_tabs_above_the_first_heading(nav_html: Callable[..., str]) -> None:
+    pages = web2md.discover_pages(nav_html())
+
+    # The leading "/docs/guides" item precedes any heading and must be dropped.
+    assert all(p.path.startswith("/style") for p in pages)
+    assert all(p.title != "Guides" for p in pages)
+
+
+def test_ignores_non_style_links_and_deduplicates_repeats(nav_html: Callable[..., str]) -> None:
+    extra = (
+        '<li class="devsite-nav-item"><a href="/docs/elsewhere">Elsewhere</a></li>'
+        '<li class="devsite-nav-item"><a href="/style/page-0-0">Duplicate</a></li>'
+        '<li class="devsite-nav-item"><a href="/style/page-0-0?hl=en#frag">Query dup</a></li>'
+        '<li class="devsite-nav-item"><span>No link</span></li>'
+    )
+    pages = web2md.discover_pages(nav_html(extra_items=extra))
+
+    assert len(pages) == 72
+    assert [p.slug for p in pages].count("page-0-0") == 1
+    assert all("elsewhere" not in p.path for p in pages)
+
+
+def test_missing_navigation_is_fatal() -> None:
+    with pytest.raises(SystemExit, match="navigation not found"):
+        web2md.discover_pages("<html><body><p>no nav here</p></body></html>")
+
+
+def test_too_few_pages_is_fatal(nav_html: Callable[..., str]) -> None:
+    with pytest.raises(SystemExit, match="unexpected page count"):
+        web2md.discover_pages(nav_html(sections=5, per_section=2))
+
+
+def test_too_many_pages_is_fatal(nav_html: Callable[..., str]) -> None:
+    with pytest.raises(SystemExit, match="unexpected page count"):
+        web2md.discover_pages(nav_html(sections=6, per_section=40))
+
+
+def test_too_few_sections_is_fatal(nav_html: Callable[..., str]) -> None:
+    with pytest.raises(SystemExit, match="unexpected section count"):
+        web2md.discover_pages(nav_html(sections=4, per_section=20))
diff --git a/web2md/tests/test_fetch.py b/web2md/tests/test_fetch.py
new file mode 100644
index 0000000..4643c98
--- /dev/null
+++ b/web2md/tests/test_fetch.py
@@ -0,0 +1,198 @@
+"""`fetch_html` — caching, retries, and backoff, all against a mock transport.
+
+`httpx.MockTransport` maps requests to canned responses without opening a socket
+(https://www.python-httpx.org/advanced/transports).
+"""
+
+from __future__ import annotations
+
+from collections.abc import Callable, Iterable
+from pathlib import Path
+
+import httpx
+import pytest
+
+import web2md
+
+URL = "https://developers.google.com/style/tense"
+
+
+def client_returning(
+    responses: Iterable[httpx.Response], log: list[httpx.Request]
+) -> httpx.Client:
+    """A client that replays `responses` in order, recording each request."""
+    queue = list(responses)
+
+    def handler(request: httpx.Request) -> httpx.Response:
+        log.append(request)
+        return queue.pop(0) if queue else httpx.Response(200, text="<html>fallback</html>")
+
+    return httpx.Client(transport=httpx.MockTransport(handler))
+
+
+@pytest.fixture
+def fetch(tmp_path: Path) -> Callable[..., tuple[str, list[httpx.Request]]]:
+    def _fetch(
+        responses: Iterable[httpx.Response],
+        *,
+        refresh: bool = False,
+        cache_name: str = "tense.html",
+    ) -> tuple[str, list[httpx.Request]]:
+        log: list[httpx.Request] = []
+        with client_returning(responses, log) as client:
+            html = web2md.fetch_html(client, URL, tmp_path / cache_name, refresh=refresh)
+        return html, log
+
+    return _fetch
+
+
+def test_warm_cache_is_served_without_a_request(tmp_path: Path) -> None:
+    cache = tmp_path / "tense.html"
+    cache.write_text("<html>cached</html>", encoding="utf-8")
+    log: list[httpx.Request] = []
+
+    with client_returning([], log) as client:
+        html = web2md.fetch_html(client, URL, cache, refresh=False)
+
+    assert html == "<html>cached</html>"
+    assert log == []
+
+
+def test_cache_miss_fetches_and_writes_the_file(
+    fetch: Callable[..., tuple[str, list[httpx.Request]]], tmp_path: Path
+) -> None:
+    html, log = fetch([httpx.Response(200, text="<html>fresh</html>")])
+
+    assert html == "<html>fresh</html>"
+    assert (tmp_path / "tense.html").read_text(encoding="utf-8") == "<html>fresh</html>"
+    assert len(log) == 1
+    assert log[0].url.params["hl"] == "en"
+
+
+def test_cache_directory_is_created_on_demand(tmp_path: Path) -> None:
+    cache = tmp_path / "nested" / "deeper" / "tense.html"
+    log: list[httpx.Request] = []
+
+    with client_returning([httpx.Response(200, text="<html>x</html>")], log) as client:
+        web2md.fetch_html(client, URL, cache, refresh=False)
+
+    assert cache.read_text(encoding="utf-8") == "<html>x</html>"
+
+
+def test_refresh_bypasses_a_warm_cache(tmp_path: Path) -> None:
+    cache = tmp_path / "tense.html"
+    cache.write_text("<html>stale</html>", encoding="utf-8")
+    log: list[httpx.Request] = []
+
+    with client_returning([httpx.Response(200, text="<html>fresh</html>")], log) as client:
+        html = web2md.fetch_html(client, URL, cache, refresh=True)
+
+    assert html == "<html>fresh</html>"
+    assert cache.read_text(encoding="utf-8") == "<html>fresh</html>"
+    assert len(log) == 1
+
+
+def test_rate_limit_is_retried_honouring_retry_after(
+    fetch: Callable[..., tuple[str, list[httpx.Request]]],
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    slept: list[float] = []
+    monkeypatch.setattr(web2md.time, "sleep", slept.append)
+
+    html, log = fetch(
+        [
+            httpx.Response(429, headers={"Retry-After": "7"}, text=""),
+            httpx.Response(200, text="<html>ok</html>"),
+        ]
+    )
+
+    assert html == "<html>ok</html>"
+    assert len(log) == 2
+    assert 7.0 in slept
+
+
+def test_non_numeric_retry_after_falls_back_to_the_backoff_delay(
+    fetch: Callable[..., tuple[str, list[httpx.Request]]],
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    slept: list[float] = []
+    monkeypatch.setattr(web2md.time, "sleep", slept.append)
+
+    html, _ = fetch(
+        [
+            httpx.Response(429, headers={"Retry-After": "Wed, 21 Oct 2026 07:28:00 GMT"}, text=""),
+            httpx.Response(200, text="<html>ok</html>"),
+        ]
+    )
+
+    assert html == "<html>ok</html>"
+    assert slept[0] == web2md.REQUEST_DELAY_S
+
+
+def test_server_errors_are_retried(
+    fetch: Callable[..., tuple[str, list[httpx.Request]]],
+) -> None:
+    html, log = fetch(
+        [
+            httpx.Response(503, text=""),
+            httpx.Response(500, text=""),
+            httpx.Response(200, text="<html>ok</html>"),
+        ]
+    )
+
+    assert html == "<html>ok</html>"
+    assert len(log) == 3
+
+
+def test_client_errors_are_retried_until_the_budget_runs_out(
+    fetch: Callable[..., tuple[str, list[httpx.Request]]],
+) -> None:
+    # A 404 skips the 429/5xx branch and reaches raise_for_status, which
+    # fetch_html catches as an HTTPError and retries like a transport failure.
+    with pytest.raises(SystemExit, match="failed to fetch"):
+        fetch([httpx.Response(404, text="") for _ in range(web2md.MAX_RETRIES)])
+
+
+def test_exhausting_the_retry_budget_is_fatal(tmp_path: Path) -> None:
+    log: list[httpx.Request] = []
+    responses = [httpx.Response(500, text="") for _ in range(web2md.MAX_RETRIES)]
+
+    with (
+        client_returning(responses, log) as client,
+        pytest.raises(SystemExit, match="failed to fetch"),
+    ):
+        web2md.fetch_html(client, URL, tmp_path / "tense.html", refresh=False)
+
+    assert len(log) == web2md.MAX_RETRIES
+    assert not (tmp_path / "tense.html").exists()
+
+
+def test_transport_errors_are_retried_then_reported(tmp_path: Path) -> None:
+    attempts = 0
+
+    def handler(request: httpx.Request) -> httpx.Response:
+        nonlocal attempts
+        attempts += 1
+        raise httpx.ConnectError("boom", request=request)
+
+    with (
+        httpx.Client(transport=httpx.MockTransport(handler)) as client,
+        pytest.raises(SystemExit, match="boom"),
+    ):
+        web2md.fetch_html(client, URL, tmp_path / "tense.html", refresh=False)
+
+    assert attempts == web2md.MAX_RETRIES
+
+
+def test_backoff_grows_and_is_capped(
+    fetch: Callable[..., tuple[str, list[httpx.Request]]],
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    slept: list[float] = []
+    monkeypatch.setattr(web2md.time, "sleep", slept.append)
+
+    with pytest.raises(SystemExit):
+        fetch([httpx.Response(500, text="") for _ in range(web2md.MAX_RETRIES)])
+
+    assert slept == [0.2, 0.4, 0.8, 1.6, 3.2]
+    assert max(slept) <= 30.0
diff --git a/web2md/tests/test_links.py b/web2md/tests/test_links.py
new file mode 100644
index 0000000..f8de780
--- /dev/null
+++ b/web2md/tests/test_links.py
@@ -0,0 +1,180 @@
+"""Internal link rewriting and anchor injection."""
+
+from __future__ import annotations
+
+from collections.abc import Callable
+
+import pytest
+from bs4 import Tag
+
+import web2md
+
+# The page under review, the book's pages keyed by path, and the anchor map.
+LinkContext = tuple[web2md.Page, dict[str, web2md.Page], dict[tuple[str, str], str]]
+
+
+@pytest.fixture
+def link_context(make_page: Callable[..., web2md.Page]) -> LinkContext:
+    """A two-page book with one known anchor on each page."""
+    here = make_page("word-list", title="Word list")
+    there = make_page("tense", title="Tense")
+    home = make_page("style", title="Overview", path="/style")
+    pages_by_path = {p.path: p for p in (here, there, home)}
+    anchor_map = {
+        ("word-list", "abbreviations"): "word-list--abbreviations",
+        ("tense", "present"): "tense--present",
+    }
+    return here, pages_by_path, anchor_map
+
+
+def test_same_page_fragment_resolves_to_the_namespaced_anchor(link_context: LinkContext) -> None:
+    here, pages_by_path, anchor_map = link_context
+    unresolved: list[str] = []
+
+    result = web2md.rewrite_internal_href(
+        "#abbreviations", here, pages_by_path, anchor_map, unresolved
+    )
+
+    assert result == "#word-list--abbreviations"
+    assert unresolved == []
+
+
+def test_unknown_same_page_fragment_falls_back_to_an_absolute_url(
+    link_context: LinkContext,
+) -> None:
+    here, pages_by_path, anchor_map = link_context
+    unresolved: list[str] = []
+
+    result = web2md.rewrite_internal_href("#missing", here, pages_by_path, anchor_map, unresolved)
+
+    assert result == "https://developers.google.com/style/word-list#missing"
+    assert len(unresolved) == 1
+
+
+def test_empty_fragment_is_returned_untouched(link_context: LinkContext) -> None:
+    here, pages_by_path, anchor_map = link_context
+
+    assert web2md.rewrite_internal_href("", here, pages_by_path, anchor_map, []) == ""
+
+
+def test_cross_page_fragment_resolves(link_context: LinkContext) -> None:
+    here, pages_by_path, anchor_map = link_context
+    unresolved: list[str] = []
+
+    result = web2md.rewrite_internal_href(
+        "/style/tense#present", here, pages_by_path, anchor_map, unresolved
+    )
+
+    assert result == "#tense--present"
+    assert unresolved == []
+
+
+def test_cross_page_fragment_on_an_unknown_page_is_absolutised(link_context: LinkContext) -> None:
+    here, pages_by_path, anchor_map = link_context
+    unresolved: list[str] = []
+
+    result = web2md.rewrite_internal_href(
+        "/style/nowhere#x", here, pages_by_path, anchor_map, unresolved
+    )
+
+    assert result == "https://developers.google.com/style/nowhere#x"
+    assert len(unresolved) == 1
+
+
+@pytest.mark.parametrize("href", ["/style/tense", "/style/tense/"])
+def test_cross_page_without_fragment_becomes_a_page_anchor(
+    link_context: LinkContext, href: str
+) -> None:
+    here, pages_by_path, anchor_map = link_context
+
+    assert web2md.rewrite_internal_href(href, here, pages_by_path, anchor_map, []) == "#tense"
+
+
+def test_book_root_link_becomes_the_overview_anchor(link_context: LinkContext) -> None:
+    here, pages_by_path, anchor_map = link_context
+
+    assert web2md.rewrite_internal_href("/style", here, pages_by_path, anchor_map, []) == "#style"
+
+
+def test_unknown_style_page_without_fragment_is_absolutised(link_context: LinkContext) -> None:
+    here, pages_by_path, anchor_map = link_context
+    unresolved: list[str] = []
+
+    result = web2md.rewrite_internal_href(
+        "/style/nowhere", here, pages_by_path, anchor_map, unresolved
+    )
+
+    assert result == "https://developers.google.com/style/nowhere"
+    assert len(unresolved) == 1
+
+
+@pytest.mark.parametrize(
+    "href",
+    [
+        "https://example.com/page",
+        "mailto:someone@example.com",
+        "relative/thing",
+    ],
+)
+def test_external_and_non_http_links_are_untouched(link_context: LinkContext, href: str) -> None:
+    here, pages_by_path, anchor_map = link_context
+
+    assert web2md.rewrite_internal_href(href, here, pages_by_path, anchor_map, []) == href
+
+
+@pytest.mark.parametrize(
+    ("href", "expected"),
+    [
+        ("/docs/something", "https://developers.google.com/docs/something"),
+        (
+            "https://developers.google.com/docs/something",
+            "https://developers.google.com/docs/something",
+        ),
+    ],
+)
+def test_non_style_google_links_are_absolutised(
+    link_context: LinkContext, href: str, expected: str
+) -> None:
+    here, pages_by_path, anchor_map = link_context
+
+    assert web2md.rewrite_internal_href(href, here, pages_by_path, anchor_map, []) == expected
+
+
+def test_apply_anchors_emits_a_shared_id_once_and_strips_it(
+    frag: Callable[[str], Tag], make_page: Callable[..., web2md.Page]
+) -> None:
+    # DevSite repeats the same id on a section and the heading inside it.
+    body = frag('<section id="terms"><h2 id="terms">Terms</h2></section>')
+    page = make_page("word-list")
+    anchor_map = {("word-list", "terms"): "word-list--terms"}
+
+    web2md.apply_anchors_and_links(body, page, {page.path: page}, anchor_map, [])
+
+    anchors = body.find_all("a", id=True)
+    assert [a["id"] for a in anchors] == ["word-list--terms"]
+    assert body.find(attrs={"id": "terms"}) is None
+
+
+def test_apply_anchors_rewrites_hrefs_and_absolutises_images(
+    frag: Callable[[str], Tag], make_page: Callable[..., web2md.Page]
+) -> None:
+    body = frag('<p><a href="#abbr">jump</a><img src="/images/x.png" alt="x"></p>')
+    page = make_page("word-list")
+    anchor_map = {("word-list", "abbr"): "word-list--abbr"}
+
+    web2md.apply_anchors_and_links(body, page, {page.path: page}, anchor_map, [])
+
+    assert body.a["href"] == "#word-list--abbr"
+    assert body.img["src"] == "https://developers.google.com/images/x.png"
+
+
+def test_apply_anchors_ignores_ids_absent_from_the_map(
+    frag: Callable[[str], Tag], make_page: Callable[..., web2md.Page]
+) -> None:
+    body = frag('<h2 id="unmapped">Heading</h2>')
+    page = make_page("word-list")
+
+    web2md.apply_anchors_and_links(body, page, {page.path: page}, {}, [])
+
+    assert body.find("a") is None
+    assert body.h2["id"] == "unmapped"
diff --git a/web2md/tests/test_run.py b/web2md/tests/test_run.py
new file mode 100644
index 0000000..309c27a
--- /dev/null
+++ b/web2md/tests/test_run.py
@@ -0,0 +1,147 @@
+"""End-to-end `run()` over a synthetic site, served by a mock transport.
+
+`run()` builds its own `httpx.Client`, so the client factory is swapped for one
+that injects `httpx.MockTransport`. Nothing here touches the network or any path
+outside `tmp_path`.
+"""
+
+from __future__ import annotations
+
+from collections.abc import Callable
+from pathlib import Path
+from typing import Any
+
+import httpx
+import pytest
+
+import web2md
+
+SECTIONS = 6
+PER_SECTION = 12
+
+# What httpx.MockTransport expects: a request in, a canned response out.
+Handler = Callable[[httpx.Request], httpx.Response]
+
+
+def _page_html(section: int, item: int) -> str:
+    return f"""<html><body>
+      <div class="devsite-article-body">
+        <section id="s{section}-{item}">
+          <h2 id="s{section}-{item}">Heading {section}.{item}</h2>
+          <p>Prose for page {section}.{item}.</p>
+          <p>A link <a href="/style/page-0-0">home</a> and
+             a fragment <a href="#s{section}-{item}">self</a>.</p>
+          <devsite-code><pre>example {section}</pre></devsite-code>
+          <aside class="note"><p>Remember this.</p></aside>
+          <devsite-feedback>rate me</devsite-feedback>
+        </section>
+      </div>
+    </body></html>"""
+
+
+@pytest.fixture
+def requested() -> list[str]:
+    """Every path the mock transport was asked for, in order."""
+    return []
+
+
+@pytest.fixture
+def site(nav_html: Callable[..., str], requested: list[str]) -> Handler:
+    """Serve the synthetic book: the nav at /style, a body at every page path."""
+    index = f"<html><body>{nav_html(sections=SECTIONS, per_section=PER_SECTION)}</body></html>"
+
+    def handler(request: httpx.Request) -> httpx.Response:
+        path = request.url.path
+        requested.append(path)
+        if path == "/style":
+            return httpx.Response(200, text=index)
+        _, section, item = path.removeprefix("/style/").split("-")
+        return httpx.Response(200, text=_page_html(int(section), int(item)))
+
+    return handler
+
+
+@pytest.fixture
+def offline(monkeypatch: pytest.MonkeyPatch, site: Handler) -> None:
+    """Point `run()`'s client at the synthetic site and relax the size guards."""
+    real_client = httpx.Client
+
+    def factory(**kwargs: Any) -> httpx.Client:
+        return real_client(transport=httpx.MockTransport(site), **kwargs)
+
+    monkeypatch.setattr(web2md.httpx, "Client", factory)
+    monkeypatch.setattr(web2md, "SIZE_MIN", 0)
+    monkeypatch.setattr(web2md, "SIZE_MAX", 10_000_000)
+    monkeypatch.setattr(web2md, "WORD_LIST_TERM_EXPECTED", 0)
+    monkeypatch.setattr(web2md, "WORD_LIST_TERM_TOLERANCE", 0)
+
+
+def test_run_writes_a_complete_document(offline: None, tmp_path: Path) -> None:
+    output = tmp_path / "out" / "guide.md"
+    cache = tmp_path / "cache"
+
+    web2md.run(refresh=False, cache_dir=cache, output=output)
+
+    md = output.read_text(encoding="utf-8")
+    assert md.startswith("---\ntype: Website\n")
+    assert "# Google Developer Documentation Style Guide" in md
+    assert "## Table of contents" in md
+    for section in range(SECTIONS):
+        assert f"\n## Section {section}\n" in md
+        for item in range(PER_SECTION):
+            assert f'<a id="page-{section}-{item}"></a>' in md
+            assert f"Prose for page {section}.{item}." in md
+
+
+def test_run_populates_the_cache_and_then_reuses_it(
+    offline: None, tmp_path: Path, requested: list[str]
+) -> None:
+    output = tmp_path / "guide.md"
+    cache = tmp_path / "cache"
+    total = SECTIONS * PER_SECTION + 1  # every page, plus the nav index
+
+    web2md.run(refresh=False, cache_dir=cache, output=output)
+
+    cached = {p.name for p in cache.iterdir()}
+    assert "_index.html" in cached
+    assert len(cached) == total
+    assert len(requested) == total
+
+    # A second run must serve everything from disk and issue no further requests.
+    requested.clear()
+    web2md.run(refresh=False, cache_dir=cache, output=output)
+
+    assert requested == []
+
+
+def test_refresh_refetches_every_page(offline: None, tmp_path: Path, requested: list[str]) -> None:
+    cache = tmp_path / "cache"
+    output = tmp_path / "guide.md"
+    web2md.run(refresh=False, cache_dir=cache, output=output)
+    requested.clear()
+
+    web2md.run(refresh=True, cache_dir=cache, output=output)
+
+    assert len(requested) == SECTIONS * PER_SECTION + 1
+
+
+def test_run_applies_the_cleaning_and_conversion_rules(offline: None, tmp_path: Path) -> None:
+    output = tmp_path / "guide.md"
+
+    web2md.run(refresh=False, cache_dir=tmp_path / "cache", output=output)
+    md = output.read_text(encoding="utf-8")
+
+    assert "rate me" not in md            # devsite-feedback dropped
+    assert "devsite-code" not in md       # unwrapped to a fenced block
+    assert "```\nexample 0\n```" in md
+    assert "> [!NOTE]\n> Remember this." in md
+    assert "#### Heading 0.0" in md       # h2 shifted down by two
+    assert "[home](#page-0-0)" in md      # cross-page link became an anchor
+
+
+def test_run_creates_missing_parent_directories(offline: None, tmp_path: Path) -> None:
+    output = tmp_path / "deeply" / "nested" / "guide.md"
+
+    web2md.run(refresh=False, cache_dir=tmp_path / "c", output=output)
+
+    assert output.is_file()
diff --git a/web2md/tests/test_slugs.py b/web2md/tests/test_slugs.py
new file mode 100644
index 0000000..1b9547d
--- /dev/null
+++ b/web2md/tests/test_slugs.py
@@ -0,0 +1,70 @@
+"""Slug, anchor, and URL helpers."""
+
+from __future__ import annotations
+
+import pytest
+from bs4 import BeautifulSoup
+
+import web2md
+
+
+@pytest.mark.parametrize(
+    ("path", "expected"),
+    [
+        ("/style", "style"),
+        ("/style/", "style"),
+        ("/style/word-list", "word-list"),
+        ("/style/tense", "tense"),
+        ("/style/a/b", "a/b"),
+    ],
+)
+def test_slug_from_path(path: str, expected: str) -> None:
+    assert web2md.slug_from_path(path) == expected
+
+
+@pytest.mark.parametrize(
+    ("raw", "expected"),
+    [
+        ("plain", "plain"),
+        ("keeps.dots_and-dashes", "keeps.dots_and-dashes"),
+        ("has space", "hasu0020space"),
+        ("café", "cafu00e9"),
+        ("a/b", "au002fb"),
+    ],
+)
+def test_sanitize_id_escapes_only_unsafe_characters(raw: str, expected: str) -> None:
+    assert web2md.sanitize_id(raw) == expected
+
+
+def test_sanitize_id_falls_back_for_empty_input() -> None:
+    assert web2md.sanitize_id("") == "id"
+
+
+def test_namespaced_anchor_prefixes_with_page_slug() -> None:
+    assert web2md.namespaced_anchor("word-list", "abbreviations") == "word-list--abbreviations"
+    assert web2md.namespaced_anchor("word-list", "a b") == "word-list--au0020b"
+
+
+@pytest.mark.parametrize(
+    ("href", "expected"),
+    [
+        ("/style/tense", "https://developers.google.com/style/tense"),
+        ("https://example.com/x", "https://example.com/x"),
+        ("//example.com/x", "https://example.com/x"),
+    ],
+)
+def test_absolutize_url(href: str, expected: str) -> None:
+    assert web2md.absolutize_url(href) == expected
+
+
+def test_nav_text_prefers_the_devsite_span() -> None:
+    el = BeautifulSoup(
+        '<a href="/style/x">ignored<span class="devsite-nav-text">Real title</span></a>',
+        "lxml",
+    ).a
+    assert web2md.nav_text(el) == "Real title"
+
+
+def test_nav_text_falls_back_to_full_text_and_unescapes() -> None:
+    el = BeautifulSoup('<a href="/style/x">Do&amp;Don&#39;t</a>', "lxml").a
+    assert web2md.nav_text(el) == "Do&Don't"
diff --git a/web2md/tests/test_validate.py b/web2md/tests/test_validate.py
new file mode 100644
index 0000000..a022763
--- /dev/null
+++ b/web2md/tests/test_validate.py
@@ -0,0 +1,153 @@
+"""`validate_output` — the OK path and every error branch.
+
+The real thresholds expect a ~0.5 MB document with ~598 word-list terms, so each
+test relaxes them via monkeypatch and then breaks exactly one rule.
+"""
+
+from __future__ import annotations
+
+from collections.abc import Callable
+
+import pytest
+
+import web2md
+
+# The discovered pages and their converted bodies, as validate_output takes them.
+Book = tuple[list[web2md.Page], dict[str, str]]
+
+
+@pytest.fixture(autouse=True)
+def relaxed_thresholds(monkeypatch: pytest.MonkeyPatch) -> None:
+    """Neutralise the size and word-list guards so fixtures can stay tiny."""
+    monkeypatch.setattr(web2md, "SIZE_MIN", 0)
+    monkeypatch.setattr(web2md, "SIZE_MAX", 10_000_000)
+    monkeypatch.setattr(web2md, "WORD_LIST_TERM_EXPECTED", 0)
+    monkeypatch.setattr(web2md, "WORD_LIST_TERM_TOLERANCE", 0)
+
+
+@pytest.fixture
+def book(make_page: Callable[..., web2md.Page]) -> Book:
+    """A one-page book whose body carries a single resolvable anchor."""
+    pages = [make_page("tense", section="Language", title="Tense")]
+    bodies = {"tense": '<a id="tense--present"></a>\n\n**present**\n\nUse the present tense.'}
+    return pages, bodies
+
+
+def _md(body: str) -> str:
+    return f"# Guide\n\n{body}\n"
+
+
+def test_valid_output_passes(book: Book, capsys: pytest.CaptureFixture[str]) -> None:
+    pages, bodies = book
+    md = _md(f'<a id="tense"></a>\n\n{bodies["tense"]}\n\n[Present](#tense--present)')
+
+    web2md.validate_output(md, pages, bodies)
+
+    assert "Validation OK" in capsys.readouterr().err
+
+
+def test_missing_page_body_is_reported(make_page: Callable[..., web2md.Page]) -> None:
+    pages = [make_page("tense"), make_page("voice")]
+
+    with pytest.raises(SystemExit):
+        web2md.validate_output(_md("text"), pages, {"tense": "body"})
+
+
+def test_broken_internal_link_is_reported(
+    book: Book, capsys: pytest.CaptureFixture[str]
+) -> None:
+    pages, bodies = book
+    md = _md('[Nowhere](#does-not-exist)')
+
+    with pytest.raises(SystemExit):
+        web2md.validate_output(md, pages, bodies)
+
+    assert "broken internal links" in capsys.readouterr().err
+
+
+def test_page_slug_anchors_count_as_valid_targets(book: Book) -> None:
+    pages, bodies = book
+
+    web2md.validate_output(_md("[Tense](#tense)"), pages, bodies)
+
+
+def test_devsite_leftovers_outside_a_fence_are_reported(
+    book: Book, capsys: pytest.CaptureFixture[str]
+) -> None:
+    pages, bodies = book
+
+    with pytest.raises(SystemExit):
+        web2md.validate_output(_md("a devsite-toc leaked through"), pages, bodies)
+
+    assert "devsite-" in capsys.readouterr().err
+
+
+def test_devsite_markup_inside_a_fence_is_allowed(book: Book) -> None:
+    pages, bodies = book
+
+    web2md.validate_output(_md("```\n<devsite-toc></devsite-toc>\n```"), pages, bodies)
+
+
+def test_pandoc_definition_markers_are_reported(
+    book: Book, capsys: pytest.CaptureFixture[str]
+) -> None:
+    pages, bodies = book
+
+    with pytest.raises(SystemExit):
+        web2md.validate_output(_md(":   a definition body"), pages, bodies)
+
+    assert "Pandoc-style definition list markers" in capsys.readouterr().err
+
+
+def test_headings_deeper_than_h6_are_reported(
+    book: Book, capsys: pytest.CaptureFixture[str]
+) -> None:
+    pages, bodies = book
+
+    with pytest.raises(SystemExit):
+        web2md.validate_output(_md("####### Seven"), pages, bodies)
+
+    assert "deeper than h6" in capsys.readouterr().err
+
+
+@pytest.mark.parametrize(("low", "high"), [(10_000_000, 20_000_000), (0, 5)])
+def test_output_size_outside_the_band_is_reported(
+    book: Book, monkeypatch: pytest.MonkeyPatch, capsys: pytest.CaptureFixture[str],
+    low: int, high: int,
+) -> None:
+    pages, bodies = book
+    monkeypatch.setattr(web2md, "SIZE_MIN", low)
+    monkeypatch.setattr(web2md, "SIZE_MAX", high)
+
+    with pytest.raises(SystemExit):
+        web2md.validate_output(_md("text"), pages, bodies)
+
+    assert "outside" in capsys.readouterr().err
+
+
+def test_word_list_term_count_is_checked(
+    make_page: Callable[..., web2md.Page],
+    monkeypatch: pytest.MonkeyPatch,
+    capsys: pytest.CaptureFixture[str],
+) -> None:
+    monkeypatch.setattr(web2md, "WORD_LIST_TERM_EXPECTED", 100)
+    monkeypatch.setattr(web2md, "WORD_LIST_TERM_TOLERANCE", 5)
+    pages = [make_page("word-list")]
+    bodies = {"word-list": '<a id="word-list--a"></a>\n\n**a**'}
+
+    with pytest.raises(SystemExit):
+        web2md.validate_output(_md(bodies["word-list"]), pages, bodies)
+
+    assert "word-list term count 1 outside 95–105" in capsys.readouterr().err
+
+
+def test_word_list_terms_within_tolerance_pass(
+    make_page: Callable[..., web2md.Page], monkeypatch: pytest.MonkeyPatch
+) -> None:
+    monkeypatch.setattr(web2md, "WORD_LIST_TERM_EXPECTED", 3)
+    monkeypatch.setattr(web2md, "WORD_LIST_TERM_TOLERANCE", 1)
+    pages = [make_page("word-list")]
+    body = "\n\n".join(f'<a id="word-list--{n}"></a>\n\n**{n}**' for n in "abc")
+    bodies = {"word-list": body}
+
+    web2md.validate_output(_md(body), pages, bodies)

From fc326680c52c700fced3c1fdf7a24ef9a2920ed8 Mon Sep 17 00:00:00 2001
From: lars20070 <l.nilse@dunelm.org.uk>
Date: Sat, 1 Aug 2026 09:58:09 +0200
Subject: [PATCH 12/15] Add SOURCE_URL and OUTPUT_FILE

---
 AGENTS.md                      |  4 +-
 README.md                      | 21 ++++++++
 web2md/README.md               | 29 ++++++++--
 web2md/src/web2md.py           | 47 +++++++++-------
 web2md/tests/test_assemble.py  |  5 +-
 web2md/tests/test_cli.py       | 10 ++--
 web2md/tests/test_constants.py | 98 ++++++++++++++++++++++++++++++++++
 7 files changed, 184 insertions(+), 30 deletions(-)
 create mode 100644 web2md/tests/test_constants.py

diff --git a/AGENTS.md b/AGENTS.md
index 0693f3a..7db2e33 100644
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -20,7 +20,9 @@ that turns a PDF into Markdown with `marker`; it is manual and not wired into
 the `make` pipeline.
 
 `web2md/` is the other upstream step: a deterministic scraper that fetches a
-website into a single file under `md/`, driven by `make scrape`. It is the only
+website into a single file under `md/`, driven by `make scrape`. Which site and
+which output filename live in two constants at the top of `web2md/src/web2md.py`
+(`SOURCE_URL`, `OUTPUT_FILE`). It is the only
 first-party Python in the repo — module in `web2md/src/`, pytest suite in
 `web2md/tests/`, gitignored HTML cache in `web2md/cache/`. There is no
 `[build-system]`: the module is run by path and pytest imports it via
diff --git a/README.md b/README.md
index 78c7b76..2bae57a 100644
--- a/README.md
+++ b/README.md
@@ -321,3 +321,24 @@ model or a cloud model through OpenRouter. Expect to check its output:
 `prettier`, `markdownlint-cli2` and `cspell` catch most of what it gets wrong,
 but none of this runs unattended. [pdf2md/README.md](pdf2md/README.md) has the
 commands.
+
+## Starting from a website
+
+When the source is a documentation site rather than a file, `make scrape` walks
+it and writes one Markdown document into `md/`. Unlike the PDF step this is
+deterministic — no model involved — and it caches the fetched HTML under
+`web2md/cache/`, so re-running is cheap and `--refresh` is what goes back to the
+network.
+
+Which book it fetches and what the result is called are two constants at the top
+of [web2md/src/web2md.py](web2md/src/web2md.py):
+
+```python
+SOURCE_URL = "https://developers.google.com/style"
+OUTPUT_FILE = "GoogleDeveloperDocumentationStyleGuide.md"
+```
+
+Everything URL-shaped in the scraper is derived from `SOURCE_URL`, so retargeting
+it is a one-line edit — though the HTML selectors and sanity thresholds describe
+this particular book and would need revisiting.
+[web2md/README.md](web2md/README.md) has the details.
diff --git a/web2md/README.md b/web2md/README.md
index 58a6b67..4793c00 100644
--- a/web2md/README.md
+++ b/web2md/README.md
@@ -9,6 +9,27 @@ guide](https://developers.google.com/style). The output is one file under
 The result is a dated snapshot of a living document. Re-run with `--refresh` to
 update it.
 
+## What it fetches, and what it writes
+
+Two constants at the top of `src/web2md.py` are the only things to edit when
+pointing the scraper somewhere else:
+
+```python
+SOURCE_URL = "https://developers.google.com/style"
+OUTPUT_FILE = "GoogleDeveloperDocumentationStyleGuide.md"
+```
+
+`SOURCE_URL` is the book's landing page. The host, the site base URL, and the
+path prefix that decides which links count as in-book (`BOOK_PATH`) are all
+derived from it, so there is no second place to keep in step. `OUTPUT_FILE` is a
+bare filename; the scraper always writes it into `md/`, and `--output`
+overrides the whole path for a one-off run.
+
+Changing `SOURCE_URL` to a different book will need the DevSite selectors
+(`NAV`, `BODY`, `DROP`) and the sanity thresholds (page count, size band, and
+the `word-list` term count in `validate_output`) revisited — they describe this
+book, not the site in general.
+
 ## Layout
 
 | Path | Contents |
@@ -44,14 +65,14 @@ uv run --group test --group web2md pytest web2md/tests # the same, directly
 The suite is fully offline: HTTP is served by `httpx.MockTransport`, so no test
 opens a socket, and the only files written go to pytest's `tmp_path`. It covers
 the pure helpers (slugs, anchors, link rewriting, the Markdown converter,
-assembly, and every `validate_output` error branch), the fetch retry and caching
-logic, and one end-to-end `run()` over a synthetic 72-page site. CI runs it in
-the `test` job.
+assembly, and every `validate_output` error branch), the constants above and
+what is derived from them, the fetch retry and caching logic, and one end-to-end
+`run()` over a synthetic 72-page site. CI runs it in the `test` job.
 
 ## Markdown linting
 
 The generated file under `md/` can be checked manually with the same tools as
-`pdf2md`:
+`pdf2md`. The filename below is the current `OUTPUT_FILE`:
 
 ```bash
 prettier --check md/GoogleDeveloperDocumentationStyleGuide.md
diff --git a/web2md/src/web2md.py b/web2md/src/web2md.py
index 6e202e7..770a8da 100644
--- a/web2md/src/web2md.py
+++ b/web2md/src/web2md.py
@@ -18,8 +18,18 @@
 from bs4 import BeautifulSoup, NavigableString, Tag
 from markdownify import ATX, MarkdownConverter
 
-BASE = "https://developers.google.com"
-BOOK = f"{BASE}/style"
+# The two knobs for this scraper: which website to fetch, and what to call the file
+# it produces under md/. Everything URL-shaped below is derived from SOURCE_URL,
+# so pointing the script at another website is a one-line change here.
+SOURCE_URL = "https://developers.google.com/style"
+OUTPUT_FILE = "GoogleDeveloperDocumentationStyleGuide.md"
+
+_SOURCE = urlparse(SOURCE_URL)
+HOST = _SOURCE.netloc                                   # developers.google.com
+BASE = f"{_SOURCE.scheme}://{HOST}"                     # https://developers.google.com
+BOOK_PATH = _SOURCE.path.rstrip("/") or "/"             # /style
+BOOK_SLUG = BOOK_PATH.strip("/").rsplit("/", 1)[-1] or "index"  # style
+
 LOCALE = {"hl": "en"}
 UA = "md2okf-web2md/0.1 (+https://github.com/lars20070/md2okf)"
 NAV = ".devsite-book-nav-wrapper"
@@ -62,7 +72,7 @@
 # web2md/ — the module itself lives one level down, in web2md/src/.
 ROOT = Path(__file__).resolve().parents[1]
 DEFAULT_CACHE = ROOT / "cache"
-DEFAULT_OUTPUT = ROOT.parent / "md" / "GoogleDeveloperDocumentationStyleGuide.md"
+DEFAULT_OUTPUT = ROOT.parent / "md" / OUTPUT_FILE
 
 _SANITIZE_RE = re.compile(r"[^A-Za-z0-9._-]")
 _MD_LINK_RE = re.compile(r"\[([^\]]*)\]\((#[^)]+)\)")
@@ -89,9 +99,9 @@ class Page:
 
 
 def slug_from_path(path: str) -> str:
-    """Derive a page slug from its URL path; the book root becomes "style"."""
-    rest = path.removeprefix("/style").strip("/")
-    return rest or "style"
+    """Derive a page slug from its URL path; the book root becomes BOOK_SLUG."""
+    rest = path.removeprefix(BOOK_PATH).strip("/")
+    return rest or BOOK_SLUG
 
 
 def sanitize_id(raw: str) -> str:
@@ -152,13 +162,13 @@ def discover_pages(html: str) -> list[Page]:
         if link is None:
             continue
         href = link["href"].split("?")[0].split("#")[0]
-        if not href.startswith("/style"):
+        if not href.startswith(BOOK_PATH):
             continue
         full = urljoin(BASE, href)
         if full in seen:
             continue
         seen.add(full)
-        path = urlparse(full).path.rstrip("/") or "/style"
+        path = urlparse(full).path.rstrip("/") or BOOK_PATH
         pages.append(
             Page(
                 section=current_section,
@@ -335,22 +345,19 @@ def rewrite_internal_href(
         unresolved.append(f"{page.url} -> {href}")
         return absolutize_url(f"{page.path}#{fragment}")
 
-    # Absolute or site-relative /style/... links.
-    if parsed.netloc and parsed.netloc != "developers.google.com":
+    # Absolute or site-relative links into the book.
+    if parsed.netloc and parsed.netloc != HOST:
         return href
     if parsed.scheme and parsed.scheme not in ("http", "https"):
         return href
 
     path = parsed.path.rstrip("/") or "/"
-    if not path.startswith("/style"):
-        if href.startswith("/") or parsed.netloc == "developers.google.com":
+    if not path.startswith(BOOK_PATH):
+        if href.startswith("/") or parsed.netloc == HOST:
             return absolutize_url(href)
         return href
 
-    target_path = path if path != "/style" else "/style"
-    target_page = pages_by_path.get(target_path)
-    if target_page is None and target_path == "/style":
-        target_page = pages_by_path.get("/style")
+    target_page = pages_by_path.get(path)
 
     if fragment:
         if target_page is None:
@@ -551,14 +558,14 @@ def assemble(pages: list[Page], bodies: dict[str, str], timestamp: str) -> str:
         "type: Website",
         'title: "Google. Google Developer Documentation Style Guide."',
         'description: "Style guide for Google developer documentation"',
-        f"resource: {BOOK}",
+        f"resource: {SOURCE_URL}",
         "tags: [guide, Google]",
         f"timestamp: {timestamp}",
         "---",
         "",
         "# Google Developer Documentation Style Guide",
         "",
-        f"*Snapshot of [{BOOK}]({BOOK}) generated {timestamp[:10]}.*",
+        f"*Snapshot of [{SOURCE_URL}]({SOURCE_URL}) generated {timestamp[:10]}.*",
         "",
         build_toc(pages),
     ]
@@ -677,7 +684,7 @@ def run(*, refresh: bool, cache_dir: Path, output: Path) -> None:
         timeout=60.0,
     ) as client:
         index_html = fetch_html(
-            client, BOOK, cache_dir / "_index.html", refresh=refresh
+            client, SOURCE_URL, cache_dir / "_index.html", refresh=refresh
         )
         pages = discover_pages(index_html)
         pages_by_path = {p.path: p for p in pages}
@@ -734,7 +741,7 @@ def run(*, refresh: bool, cache_dir: Path, output: Path) -> None:
 def main(argv: list[str] | None = None) -> None:
     """Parse the command line and run the scrape."""
     parser = argparse.ArgumentParser(
-        description="Fetch developers.google.com/style into one Markdown file."
+        description=f"Fetch {SOURCE_URL} into one Markdown file."
     )
     parser.add_argument(
         "--refresh",
diff --git a/web2md/tests/test_assemble.py b/web2md/tests/test_assemble.py
index 4aee898..4e636a1 100644
--- a/web2md/tests/test_assemble.py
+++ b/web2md/tests/test_assemble.py
@@ -49,9 +49,10 @@ def test_assemble_writes_frontmatter_and_timestamp(
     assert lines[0] == "---"
     assert "type: Website" in lines
     assert f"timestamp: {TIMESTAMP}" in lines
-    assert f"resource: {web2md.BOOK}" in lines
+    assert f"resource: {web2md.SOURCE_URL}" in lines
     assert "# Google Developer Documentation Style Guide" in lines
-    assert f"*Snapshot of [{web2md.BOOK}]({web2md.BOOK}) generated 2026-08-01.*" in lines
+    snapshot = f"*Snapshot of [{web2md.SOURCE_URL}]({web2md.SOURCE_URL})"
+    assert f"{snapshot} generated 2026-08-01.*" in lines
 
 
 def test_assemble_emits_each_section_once_with_page_anchors(
diff --git a/web2md/tests/test_cli.py b/web2md/tests/test_cli.py
index 9f5e00e..47943fa 100644
--- a/web2md/tests/test_cli.py
+++ b/web2md/tests/test_cli.py
@@ -19,9 +19,13 @@ def test_default_cache_sits_beside_the_src_directory() -> None:
 
 
 def test_default_output_goes_to_the_repo_level_md_directory() -> None:
-    assert web2md.DEFAULT_OUTPUT == (
-        REPO_ROOT / "md" / "GoogleDeveloperDocumentationStyleGuide.md"
-    )
+    assert web2md.DEFAULT_OUTPUT == REPO_ROOT / "md" / web2md.OUTPUT_FILE
+
+
+def test_output_file_is_a_bare_filename() -> None:
+    # DEFAULT_OUTPUT joins it onto md/, so a path here would escape that folder.
+    assert "/" not in web2md.OUTPUT_FILE
+    assert web2md.OUTPUT_FILE.endswith(".md")
 
 
 @pytest.fixture
diff --git a/web2md/tests/test_constants.py b/web2md/tests/test_constants.py
new file mode 100644
index 0000000..3e6b46a
--- /dev/null
+++ b/web2md/tests/test_constants.py
@@ -0,0 +1,98 @@
+"""The two top-of-file knobs, and everything derived from them.
+
+SOURCE_URL and OUTPUT_FILE are the only values meant to be edited when pointing
+the scraper at a different book. These tests pin the derivation so a change to
+SOURCE_URL cannot leave a stale host or book path behind somewhere in the module.
+"""
+
+from __future__ import annotations
+
+from typing import Any
+from urllib.parse import urlparse
+
+import pytest
+
+import web2md
+
+
+def test_source_url_and_output_file_are_the_current_book() -> None:
+    assert web2md.SOURCE_URL == "https://developers.google.com/style"
+    assert web2md.OUTPUT_FILE == "GoogleDeveloperDocumentationStyleGuide.md"
+
+
+def test_host_base_and_book_path_derive_from_source_url() -> None:
+    assert web2md.HOST == "developers.google.com"
+    assert web2md.BASE == "https://developers.google.com"
+    assert web2md.BOOK_PATH == "/style"
+    assert web2md.BOOK_SLUG == "style"
+
+
+def test_book_path_is_the_source_url_path() -> None:
+    assert f"{web2md.BASE}{web2md.BOOK_PATH}" == web2md.SOURCE_URL
+
+
+def _derive(source_url: str) -> dict[str, Any]:
+    """Re-run the module's top-level derivation for an alternative SOURCE_URL."""
+    parsed = urlparse(source_url)
+    host = parsed.netloc
+    book_path = parsed.path.rstrip("/") or "/"
+    return {
+        "HOST": host,
+        "BASE": f"{parsed.scheme}://{host}",
+        "BOOK_PATH": book_path,
+        "BOOK_SLUG": book_path.strip("/").rsplit("/", 1)[-1] or "index",
+    }
+
+
+def test_derivation_matches_the_module_for_the_configured_url() -> None:
+    derived = _derive(web2md.SOURCE_URL)
+
+    assert derived["HOST"] == web2md.HOST
+    assert derived["BASE"] == web2md.BASE
+    assert derived["BOOK_PATH"] == web2md.BOOK_PATH
+    assert derived["BOOK_SLUG"] == web2md.BOOK_SLUG
+
+
+@pytest.mark.parametrize(
+    ("source_url", "host", "base", "book_path", "book_slug"),
+    [
+        (
+            "https://developers.google.com/style",
+            "developers.google.com",
+            "https://developers.google.com",
+            "/style",
+            "style",
+        ),
+        (
+            "https://example.org/docs/handbook/",
+            "example.org",
+            "https://example.org",
+            "/docs/handbook",
+            "handbook",
+        ),
+        ("http://example.org", "example.org", "http://example.org", "/", "index"),
+    ],
+)
+def test_derivation_handles_other_shapes_of_source_url(
+    source_url: str, host: str, base: str, book_path: str, book_slug: str
+) -> None:
+    derived = _derive(source_url)
+
+    assert derived["HOST"] == host
+    assert derived["BASE"] == base
+    assert derived["BOOK_PATH"] == book_path
+    assert derived["BOOK_SLUG"] == book_slug
+
+
+def test_slug_from_path_uses_book_path_and_book_slug(
+    monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    monkeypatch.setattr(web2md, "BOOK_PATH", "/docs/handbook")
+    monkeypatch.setattr(web2md, "BOOK_SLUG", "handbook")
+
+    assert web2md.slug_from_path("/docs/handbook") == "handbook"
+    assert web2md.slug_from_path("/docs/handbook/intro") == "intro"
+
+
+def test_default_output_is_named_by_output_file() -> None:
+    assert web2md.DEFAULT_OUTPUT.name == web2md.OUTPUT_FILE

From 24d22e619b921aa5a8d659b9f7f07c9f1951fc4c Mon Sep 17 00:00:00 2001
From: lars20070 <l.nilse@dunelm.org.uk>
Date: Sat, 1 Aug 2026 10:12:54 +0200
Subject: [PATCH 13/15] Rename Google style guide md file

---
 .cursor/plans/google-style-guide-to-markdown.md      |  8 ++++----
 README.md                                            |  2 +-
 ...ocumentationStyleGuide.md => GoogleStyleGuide.md} |  0
 web2md/README.md                                     | 12 ++++++------
 web2md/src/web2md.py                                 |  2 +-
 web2md/tests/test_constants.py                       |  2 +-
 6 files changed, 13 insertions(+), 13 deletions(-)
 rename md/{GoogleDeveloperDocumentationStyleGuide.md => GoogleStyleGuide.md} (100%)

diff --git a/.cursor/plans/google-style-guide-to-markdown.md b/.cursor/plans/google-style-guide-to-markdown.md
index 3772d65..bef274e 100644
--- a/.cursor/plans/google-style-guide-to-markdown.md
+++ b/.cursor/plans/google-style-guide-to-markdown.md
@@ -132,7 +132,7 @@ place to confirm it.
 
 ## Decisions
 
-- **One output file**, `md/GoogleDeveloperDocumentationStyleGuide.md`, matching
+- **One output file**, `md/GoogleStyleGuide.md`, matching
   the existing `md/TheEconomistStyleGuide2023.md` naming.
 - **New `web2md/` directory**, mirroring `pdf2md/`: one Python module plus a
   README, wired into `make lint` and a `make` target.
@@ -296,9 +296,9 @@ In-script, exiting non-zero on failure:
 Then, by hand on the host, the `pdf2md` chain:
 
 ```bash
-prettier --check md/GoogleDeveloperDocumentationStyleGuide.md
-markdownlint-cli2 md/GoogleDeveloperDocumentationStyleGuide.md
-cspell md/GoogleDeveloperDocumentationStyleGuide.md
+prettier --check md/GoogleStyleGuide.md
+markdownlint-cli2 md/GoogleStyleGuide.md
+cspell md/GoogleStyleGuide.md
 ```
 
 Finish with `make lint`, and a `make wiki-sandbox` run to confirm the file works
diff --git a/README.md b/README.md
index 2bae57a..80cbf35 100644
--- a/README.md
+++ b/README.md
@@ -335,7 +335,7 @@ of [web2md/src/web2md.py](web2md/src/web2md.py):
 
 ```python
 SOURCE_URL = "https://developers.google.com/style"
-OUTPUT_FILE = "GoogleDeveloperDocumentationStyleGuide.md"
+OUTPUT_FILE = "GoogleStyleGuide.md"
 ```
 
 Everything URL-shaped in the scraper is derived from `SOURCE_URL`, so retargeting
diff --git a/md/GoogleDeveloperDocumentationStyleGuide.md b/md/GoogleStyleGuide.md
similarity index 100%
rename from md/GoogleDeveloperDocumentationStyleGuide.md
rename to md/GoogleStyleGuide.md
diff --git a/web2md/README.md b/web2md/README.md
index 4793c00..6c7f0f4 100644
--- a/web2md/README.md
+++ b/web2md/README.md
@@ -16,7 +16,7 @@ pointing the scraper somewhere else:
 
 ```python
 SOURCE_URL = "https://developers.google.com/style"
-OUTPUT_FILE = "GoogleDeveloperDocumentationStyleGuide.md"
+OUTPUT_FILE = "GoogleStyleGuide.md"
 ```
 
 `SOURCE_URL` is the book's landing page. The host, the site base URL, and the
@@ -75,11 +75,11 @@ The generated file under `md/` can be checked manually with the same tools as
 `pdf2md`. The filename below is the current `OUTPUT_FILE`:
 
 ```bash
-prettier --check md/GoogleDeveloperDocumentationStyleGuide.md
-prettier --write md/GoogleDeveloperDocumentationStyleGuide.md  # Edits in place!
+prettier --check md/GoogleStyleGuide.md
+prettier --write md/GoogleStyleGuide.md  # Edits in place!
 
-markdownlint-cli2 md/GoogleDeveloperDocumentationStyleGuide.md
-markdownlint-cli2 --fix md/GoogleDeveloperDocumentationStyleGuide.md  # Edits in place!
+markdownlint-cli2 md/GoogleStyleGuide.md
+markdownlint-cli2 --fix md/GoogleStyleGuide.md  # Edits in place!
 
-cspell md/GoogleDeveloperDocumentationStyleGuide.md
+cspell md/GoogleStyleGuide.md
 ```
diff --git a/web2md/src/web2md.py b/web2md/src/web2md.py
index 770a8da..4cf023b 100644
--- a/web2md/src/web2md.py
+++ b/web2md/src/web2md.py
@@ -22,7 +22,7 @@
 # it produces under md/. Everything URL-shaped below is derived from SOURCE_URL,
 # so pointing the script at another website is a one-line change here.
 SOURCE_URL = "https://developers.google.com/style"
-OUTPUT_FILE = "GoogleDeveloperDocumentationStyleGuide.md"
+OUTPUT_FILE = "GoogleStyleGuide.md"
 
 _SOURCE = urlparse(SOURCE_URL)
 HOST = _SOURCE.netloc                                   # developers.google.com
diff --git a/web2md/tests/test_constants.py b/web2md/tests/test_constants.py
index 3e6b46a..5b3af0a 100644
--- a/web2md/tests/test_constants.py
+++ b/web2md/tests/test_constants.py
@@ -17,7 +17,7 @@
 
 def test_source_url_and_output_file_are_the_current_book() -> None:
     assert web2md.SOURCE_URL == "https://developers.google.com/style"
-    assert web2md.OUTPUT_FILE == "GoogleDeveloperDocumentationStyleGuide.md"
+    assert web2md.OUTPUT_FILE == "GoogleStyleGuide.md"
 
 
 def test_host_base_and_book_path_derive_from_source_url() -> None:

From 3337ee031ef07acc74878b54372a5b7887452f29 Mon Sep 17 00:00:00 2001
From: lars20070 <l.nilse@dunelm.org.uk>
Date: Sat, 1 Aug 2026 10:20:40 +0200
Subject: [PATCH 14/15] Silence `latest` warning by Coderabbit

---
 .coderabbit.yaml | 1 +
 1 file changed, 1 insertion(+)

diff --git a/.coderabbit.yaml b/.coderabbit.yaml
index cc2fe81..6be91d4 100644
--- a/.coderabbit.yaml
+++ b/.coderabbit.yaml
@@ -35,6 +35,7 @@ reviews:
     - "!md/**"
     - "!okf/**/*.md"
     - "!pdf2md/**"
+    - "!.cursor/mcp.json"
   path_instructions:
     - path: "**/*.sh"
       instructions: |

From 913aa475cddd0c36cd7cd1c79ea6f3420f5bead9 Mon Sep 17 00:00:00 2001
From: lars20070 <l.nilse@dunelm.org.uk>
Date: Sat, 1 Aug 2026 10:21:44 +0200
Subject: [PATCH 15/15] Delete plans

---
 .claude/plans/compile-wiki-skill-container.md | 413 ------------------
 .claude/plans/observable-wiki-compilation.md  | 162 -------
 .claude/plans/observe-pi-agent-sandbox.md     | 148 -------
 .../plans/google-style-guide-to-markdown.md   | 325 --------------
 .../plans/google_style_guide_scraper_plan.md  | 313 -------------
 5 files changed, 1361 deletions(-)
 delete mode 100644 .claude/plans/compile-wiki-skill-container.md
 delete mode 100644 .claude/plans/observable-wiki-compilation.md
 delete mode 100644 .claude/plans/observe-pi-agent-sandbox.md
 delete mode 100644 .cursor/plans/google-style-guide-to-markdown.md
 delete mode 100644 .cursor/plans/google_style_guide_scraper_plan.md

diff --git a/.claude/plans/compile-wiki-skill-container.md b/.claude/plans/compile-wiki-skill-container.md
deleted file mode 100644
index 00acf05..0000000
--- a/.claude/plans/compile-wiki-skill-container.md
+++ /dev/null
@@ -1,413 +0,0 @@
-# Refactor the container agent's instructions into a `compile-wiki` skill
-
-## Context
-
-The sandbox runtime already works this way. Commit `4735d60` split
-[pi/sandbox/files/home/.pi/agent/AGENTS.md](pi/sandbox/files/home/.pi/agent/AGENTS.md)
-into a task-agnostic constitution (identity, workspace boundaries, OKF format
-conventions) plus a `compile-wiki` skill holding the **procedure**. Adding a
-future task ("consolidate the wiki") becomes a new skill directory rather than
-more rules in an always-on instruction file.
-
-The container runtime never got that treatment.
-[pi/container/agent/AGENTS.md](pi/container/agent/AGENTS.md) is still one flat
-100-line file mixing identity, conventions and workflow, and
-`pi/container/agent/skills/compile-wiki/SKILL.md` exists as a **0-byte
-untracked stub**. This plan fills that stub and slims AGENTS.md to match the
-sandbox's shape.
-
-**The constraint that shapes everything below:**
-[scripts/compile-wiki-container.sh:52](scripts/compile-wiki-container.sh#L52)
-passes `-xt bash`, which excludes the bash tool outright. (The sandbox dropped
-that flag in commit `6318bdd` precisely so its agent could run a linter.) The
-container agent therefore cannot run `okf-lint`, `cat SPEC.md`, or `date +%F`.
-`-xt bash` stays; the container skill is written to live within it.
-
-**`read` survives `-xt bash`.** Verified against the installed Pi package rather
-than assumed: `-xt` is a per-tool-name denylist, and `read` is a built-in
-separate from `bash` — the CLI tagline is "AI coding assistant with read, bash,
-edit, write tools", and `docs/extensions.md:1847` enumerates the built-ins as
-`read`, `bash`, `edit`, `write`, `grep`, `find`, `ls`. The `read` tool takes
-`{ path, offset?, limit? }` (`docs/extensions.md:709`), so the agent can open
-`SPEC.md` directly **and** read long sources in ranges. Losing bash costs the
-agent `okf-lint` and `date`, not file access.
-
-**Keeping `read` is also what keeps skills working.** In
-`dist/core/system-prompt.js`, the skills section is appended only under
-`if (hasRead && skills.length > 0)` — Pi advertises a skill in the system prompt
-**only when the `read` tool is available**, since a skill is useless if the model
-cannot open its `SKILL.md`. `-xt bash` leaves `read` intact, so `compile-wiki`
-will be advertised normally. The rule this implies for the future: never add
-`read` to the `-xt` list, or the skill silently vanishes from the prompt.
-
-Consequences, all confirmed with the user:
-
-- **No `okf-lint` anywhere** in the container skill or AGENTS.md — the whole
-  "Check your output with `okf-lint`" section and the `lint-okf.sh` wrapper are
-  omitted. The `SPEC.md` references stay: `SPEC.md` is baked into the image at
-  `/workspace/SPEC.md` ([pi/container/Dockerfile:26](pi/container/Dockerfile#L26))
-  and read with the `read` tool, not fetched.
-- **A `log.md`, but with no dates.** The agent writes `okf/log.md` and prepends
-  each run's entries so the newest are first — but the entries carry no
-  datestamps, because on the pinned Pi version the container agent has **no way
-  to learn the date**. Verified in the 0.82.1 tarball, not assumed:
-  `buildSystemPrompt` in `dist/core/system-prompt.js` appends only
-  `Current working directory` — no date. It used to append
-  `Current date: YYYY-MM-DD`, but release **0.80.7** (2026-07-14) removed it:
-  *"Fixed system prompt cache invalidation across dates by removing the current
-  date from the default prompt"* (CHANGELOG, issue \#6621). Both runtimes pin
-  0.82.1 ([Dockerfile](pi/container/Dockerfile), [spec.yaml](pi/sandbox/spec.yaml)),
-  so both are past that change. Without bash there is no `date +%F` either, so an
-  agent told to datestamp would have to invent the date — the exact failure the
-  "never take a version from memory" rule exists to prevent. A dateless log is
-  the honest form. (This also explains why the sandbox skill's `date +%F` call is
-  load-bearing rather than belt-and-braces.)
-- **No `scripts/` directory** in the container skill — nothing to wrap.
-
-### The dateless log must carry no `##` headings
-
-This constraint is not stylistic — it is what keeps the result lint-clean, and it
-was read out of the pinned `okf-lint@0.1.0` source rather than guessed:
-
-- **`log-date-format` is severity `error`**, and it fires on **every** `##`
-  heading in `log.md` that is not an ISO date — not merely on headings that look
-  like dates. A heading such as `## Chapter 3` produces
-  *"Log entry headings must use ISO 8601 `YYYY-MM-DD` form"*. It is **not**
-  disabled in [okf/.okflintrc.json](okf/.okflintrc.json), so it is live.
-- If `log.md` contains **no `##` headings at all**, the rule collects nothing and
-  stays silent. `log-date-order` likewise filters to ISO-dated headings, so it is
-  inert too.
-- `recommended-log` only checks that `log.md` **exists**; it never inspects the
-  format.
-
-So the container's log is a **flat bullet list under a single `#` heading**, with
-no sub-headings, newest entries prepended at the top.
-
-#### Worked example
-
-This is the target shape, written out at realistic length. It shows the state
-after **three** compile runs against the Economist Style Guide source, so the
-accumulation is visible. Reproduce this example in `AGENTS.md` so the agent has a
-concrete model rather than a description:
-
-```markdown
-# Update Log
-
-* **Update**: Expanded [8. What's in a name](/part-2/8-whats-in-a-name.md) with
-  the sections on place names, corporate names, and the treatment of titles.
-* **Creation**: Added
-  [9. American and British English](/part-2/9-american-and-british-english.md),
-  covering spelling, vocabulary and punctuation differences.
-* **Deprecation**: Folded the standalone `/part-2/hyphens.md` into
-  [7. Punctuation, mechanics and conventions](/part-2/7-punctuation-mechanics-conventions.md)
-  and removed its entry from [the Part 2 index](/part-2/index.md).
-* **Update**: Regenerated [the Part 2 index](/part-2/index.md) so it lists the
-  chapters above.
-* **Creation**: Added [Part 2](/part-2/index.md) with chapters
-  [6. Confusables and cuttables](/part-2/6-confusables-and-cuttables.md) and
-  [10. Reference](/part-2/10-reference.md).
-* **Creation**: Added [the Glossary](/glossary/index.md) with 18 grammatical
-  term definitions, among them [verb](/glossary/verb.md),
-  [noun phrase](/glossary/noun-phrase.md) and
-  [subordinate clause](/glossary/subordinate-clause.md).
-* **Update**: Added the glossary to [the root index](/index.md).
-* **Creation**: Added [Part 1](/part-1/index.md) with chapters
-  [1. Old and short: words](/part-1/1-old-and-short-words.md) through
-  [5. Editing](/part-1/5-editing.md).
-* **Creation**: Added [the Introduction](/introduction.md), presenting Orwell's
-  six rules and the principles of clarity, concision, honesty, humility and
-  lucidity.
-* **Creation**: Initialised the wiki from `md/TheEconomistStyleGuide2023.md` and
-  added [the root index](/index.md) declaring `okf_version` 0.1.
-```
-
-How to read it, and the points the agent must copy:
-
-- **One `#` heading, no `##` anywhere.** Every entry is a top-level `*` bullet.
-  Runs are not separated by headings, rules, or blank-line groups.
-- **Newest first.** The bottom four bullets are the first run (initialisation,
-  Introduction, Part 1, glossary); the middle are the second; the top two are the
-  most recent. Each run **prepends** directly beneath `# Update Log` and leaves
-  everything below untouched.
-- **No dates, anywhere** — not in headings, not inline in the prose.
-- **One bullet per page created or updated**, plus a bullet for an index
-  regenerated as a consequence. A run that touches five pages adds roughly five
-  bullets, not one summary bullet.
-- **Wrapped bullets continue with a two-space indent**, as in the
-  `8. What's in a name` and `Deprecation` entries.
-- **Links are bundle-absolute and point at a real `.md` file** —
-  `/part-2/index.md` rather than `/part-2/`. This is a robustness choice, not a
-  lint requirement: `valid-links` skips any target that does not end in `.md`
-  (`if (!stripped.toLowerCase().endsWith('.md')) continue`), so a directory-style
-  link like `/glossary/` is never checked at all. It will not be reported — but
-  it also gets no protection, so a renamed directory leaves a silently dead link.
-  Linking to `/glossary/index.md` puts the link under the rule's coverage.
-  Verified: the sandbox's existing `okf/log.md` uses the directory form and
-  `make lint-okf` reports `✓ No problems found.`
-- **The leading bold word carries the change type** (`**Creation**`,
-  `**Update**`, `**Deprecation**`) — a convention, not a requirement, but it is
-  the only signal of what kind of change an entry records once dates are gone.
-
-Note that "newest first" cannot be machine-checked here — with no dated headings
-`log-date-order` has nothing to compare — so it rests entirely on the instruction
-to prepend, and on the two-run check in Verification below.
-
-### One spec tension to state openly
-
-`SPEC.md` §7 describes the log as *"a flat list of date-grouped entries, newest
-first"*, and §9.3 makes conformance depend on `log.md` following §7 "when
-present". A dateless log keeps the flat-list-newest-first shape but drops the
-date grouping, so it departs from the format §7 illustrates. The normative `MUST`
-in §7 constrains only the *form of date headings* (`YYYY-MM-DD`), and with no
-headings at all nothing violates it — which is why the linter passes.
-
-This matters for instruction design, not just pedantry: `AGENTS.md` tells the
-agent **"where the spec and these conventions disagree, the spec wins"**. If
-AGENTS.md simply said "write a log without dates", a careful agent reading §7
-could reconcile the conflict by inventing dates. So the container's log section
-must **name the exception explicitly** — see change 1 below. This is your call
-and is implemented as asked; it is flagged only so the instructions do not fight
-each other.
-
-## Changes
-
-### 1. `pi/container/agent/AGENTS.md` — slim to shared invariants
-
-Keep, with wording unchanged where possible: *The OKF specification is the
-source of truth*, *Workspace boundaries*, *OKF wiki conventions* (structure,
-`okf_version` frontmatter on `okf/index.md`, per-directory `index.md` link
-lists, content-page frontmatter, kebab-case slugs, idempotency + index
-regeneration).
-
-Four edits:
-
-- **Core identity** — reword from "reads source Markdown … one document at a
-  time" to a maintainer of the wiki under `okf/` that receives one task per run,
-  mirroring the sandbox copy.
-- **Add a "Your task arrives as a skill" section**, modelled on the sandbox's:
-  skills live under `~/.pi/agent/skills/`, one directory per task, each with a
-  `SKILL.md`; when a run names one, read its `SKILL.md` first and follow it;
-  available skills: `compile-wiki`. Note that `AGENTS.md` gives the conventions,
-  the skill gives the procedure.
-- **Workspace boundaries** — drop the "You are invoked once per source document"
-  bullet; that is compile-specific and moves into the skill's intro.
-- **Remove the `### Fidelity` section** — it moves verbatim into the skill.
-
-Two fixes on top of the split:
-
-- **Fix the `cat SPEC.md` fence — it cannot work under `-xt bash`.** This is the
-  highest-value edit in the file: the agent's output is only as good as its grasp
-  of the spec, and today the one instruction that loads the spec names a tool the
-  container does not have. Silent failure here degrades every page it writes.
-  Replace the fenced block with an explicit `read`-tool instruction — spell out
-  the tool by name rather than saying "read the file":
-
-  > **At the start of every run, read `SPEC.md` before writing anything.** Open
-  > it with your **`read` tool**, path `SPEC.md` at the workspace root (absolute:
-  > `/workspace/SPEC.md`). Do **not** try to `cat` it — this runtime has no
-  > `bash` tool. `SPEC.md` is ~460 lines and fits in a **single `read` call**;
-  > read it in full rather than skimming. If you ever do need it in pieces, the
-  > `read` tool takes `offset` and `limit`.
-
-  Keep the surrounding rules verbatim: take the version from the spec's
-  "Versioning" section (§11, `SPEC.md:391`) for `okf_version`; where spec and
-  these conventions disagree, the spec wins; if `SPEC.md` is missing or
-  unreadable, leave the `okf_version` in `okf/index.md` unchanged and say so in
-  the final message.
-- **Add the bundle-absolute-links bullet** to `### Structure`, ported verbatim
-  from [the sandbox copy](pi/sandbox/files/home/.pi/agent/AGENTS.md): links are
-  rooted at the wiki root (`/glossary/verb.md`, not `glossary/verb.md`) and only
-  ever point at a page that exists on disk right now.
-
-Then add the log convention — new to the container, and deliberately **not** a
-verbatim port of the sandbox's:
-
-- **`### Structure`** gains the `okf/log.md` bullet: the bundle root carries
-  `log.md`, the update log; keep it present and current.
-- **A new `### Update log` section**, adapted from the sandbox's but dateless.
-  It must state four things, and must name the spec exception outright so the
-  "spec wins" rule cannot push the agent into inventing dates:
-
-  > The bundle root carries `okf/log.md`, a record of what each run changed.
-  > `index.md` and `log.md` are reserved filenames — `log.md` carries **no** YAML
-  > frontmatter and is never listed as a content entry in an index.
-  >
-  > - The log is a **flat bullet list under a single `# Update Log` heading**,
-  >   **newest first**. Every run **prepends** its entries directly beneath that
-  >   heading; never rewrite or reorder entries from earlier runs.
-  > - **Write no dates and no `##` headings.** This runtime cannot determine the
-  >   current date — it has no `bash` tool — and a guessed date is worse than
-  >   none. This **overrides** the date-grouped example in `SPEC.md` §7: it is
-  >   the one place where these conventions deliberately depart from the spec's
-  >   illustration. Do not add date headings to reconcile them.
-  > - Entries are short prose. The leading bold word (`**Creation**`,
-  >   `**Update**`, `**Deprecation**`) is a convention, not a requirement. Links
-  >   inside entries follow the same rules as everywhere else: bundle-absolute,
-  >   and pointing only at pages that exist on disk.
-  > - If `okf/log.md` does not exist yet, create it in this run.
-
-  Include the worked example from the Context section above so the shape is
-  unambiguous.
-
-### 2. `pi/container/agent/skills/compile-wiki/SKILL.md` — fill the empty stub
-
-Frontmatter matching the sandbox's (`name` must be lowercase/digits/hyphens;
-`description` says *when* to use the skill):
-
-```yaml
----
-name: compile-wiki
-description: Compile one Markdown source document from md/ into the OKF wiki under okf/. Use when a run asks you to create or update the wiki from a source document.
----
-```
-
-Body — same structure as
-[the sandbox SKILL.md](pi/sandbox/files/home/.pi/agent/skills/compile-wiki/SKILL.md),
-minus the bash and lint steps:
-
-- **Intro** — translate one document under `md/` into OKF pages under `okf/`;
-  invoked once per source document; integrate without disturbing unrelated
-  pages. Note that the conventions in `AGENTS.md` apply and are not restated —
-  keep `update log format` in that list of deferred conventions, as the sandbox
-  does, since the container now has one.
-- **Procedure**, five steps. Step 1 carries the same explicit `read`-tool
-  wording as AGENTS.md — the sandbox's step 1 is a `cat SPEC.md` fence, and
-  copying it across would hand the container agent a command it cannot run:
-
-  > 1. **Read `SPEC.md`** at the workspace root before writing anything, using
-  >    your **`read` tool** (there is no `bash` tool here, so `cat` is not an
-  >    option). Read it in full — it is ~460 lines — and follow the revision you
-  >    read.
-
-  Then: read the source document named in the prompt, read-only under `md/` →
-  survey the existing wiki for pages already covering these topics and update in
-  place → write the content pages by topic with the frontmatter `AGENTS.md`
-  requires, in bounded chunks → regenerate the affected `index.md` link lists.
-
-  **Step 6 is the log**, replacing the sandbox's `date +%F` step:
-
-  > 6. **Prepend to `okf/log.md`**, one entry per page you created or updated,
-  >    inserted directly beneath the `# Update Log` heading so the newest entries
-  >    come first. Write **no date and no `##` heading** — see "Update log" in
-  >    `AGENTS.md`. Create the file if it does not exist.
-
-- **"Write in bounded chunks"** — port verbatim. This matters *more* here than
-  in the sandbox: a truncated `write` call loses all its content.
-- **"Fidelity"** — port verbatim, moved out of AGENTS.md.
-
-Omit entirely: the `date +%F` call, the lint step, and the
-"Check your output with `okf-lint`" section.
-
-### 3. `scripts/compile-wiki-container.sh` — name the skill in the prompt
-
-Without this the refactor is inert: Pi's print mode passes `-p` verbatim, so
-there is no slash-command expansion and models do not reliably load a skill from
-its `description` alone. Follow the sandbox driver's phrasing, using the
-container's in-image path (compose sets `HOME: /home/node`, so `~` resolves, but
-prefer the explicit path for clarity):
-
-```bash
-docker compose -f "${compose_file}" run --rm -T pi \
-	-xt bash \
-	-p "Load the compile-wiki skill: read /home/node/.pi/agent/skills/compile-wiki/SKILL.md, then follow it to compile ${document_inside} into the OKF wiki under okf/." \
-	</dev/null
-```
-
-Keep `-xt bash` and `</dev/null` exactly as they are. Update the header comment
-to say the workflow now lives in the skill.
-
-### 4. Repo wiring
-
-- **Delete** `pi/container/agent/skills/.gitkeep` — the directory now has real
-  content, matching what the sandbox refactor did.
-- **`git add`** the previously untracked
-  `pi/container/agent/skills/compile-wiki/`.
-- **No [compose.yaml](pi/container/compose.yaml) change.** The skills mount
-  already exists at
-  [compose.yaml:49](pi/container/compose.yaml#L49) (`./agent/skills:/home/node/.pi/agent/skills:ro`),
-  so the new file is picked up on the next run with no rebuild.
-- **No [Makefile](Makefile) change.** markdownlint already globs
-  `pi/container/agent/**/*.md`, so `SKILL.md` is covered; the container skill
-  ships no shell scripts, so the shellcheck glob stays sandbox-only.
-- **No [Dockerfile](pi/container/Dockerfile) change.** `okf-lint` stays
-  installed in the image — the agent is simply never told to use it. Ripping the
-  install out is a separate call, not part of this refactor.
-- **[README.md](README.md)** — the section at line 121 is titled
-  "Skills (sandbox only)"; retitle it and add a short paragraph that the
-  container carries its own copy under `pi/container/agent/skills/`, delivered
-  by bind mount (so edits apply on the next run with no rebuild, unlike the
-  sandbox's copy-at-build), and that the container skill has no lint step
-  because its driver runs `-xt bash`. Also amend the "Inside the runtimes"
-  bullet at line 198 so it no longer implies both runtimes lint.
-
-## Deliberately out of scope
-
-- **Removing `-xt bash`** from the container driver. That is the premise of this
-  design, not an obstacle to it.
-- **Changing the sandbox's log format.** The sandbox keeps its dated,
-  `##`-grouped log — it has bash and `date +%F`, so it can. The two runtimes will
-  therefore produce **different log shapes** from the same wiki. That is a
-  deliberate consequence of the container's tool restriction, not drift to be
-  reconciled later. Worth knowing before anyone runs both drivers against one
-  `okf/`: a dated sandbox run and a dateless container run would interleave two
-  formats in one file, and the sandbox's `##` date headings would remain valid
-  while the container's flat entries sit above or below them.
-- **`pi/container/agent/models.json`** — the `models` array that capped output at
-  16,384 tokens is **already removed** in your working tree, matching the
-  sandbox. Worth folding in while you are there: the file currently ends without
-  a trailing newline.
-
-## Verification
-
-1. `make lint` — markdownlint over the new `SKILL.md` and the edited
-   `AGENTS.md`, shellcheck over the edited driver.
-2. `make validate` — required by [AGENTS.md](AGENTS.md) for any change under
-   `pi/` or `scripts/`. Nothing under `pi/sandbox/` changes, so it should stay
-   `VALID`.
-3. Confirm the skill reaches the container (needs `OPENROUTER_API_KEY` set, but
-   makes no model call):
-
-   ```bash
-   docker compose -f pi/container/compose.yaml build
-   docker compose -f pi/container/compose.yaml run --rm -T --entrypoint sh pi \
-     -c 'ls -l /home/node/.pi/agent/skills/compile-wiki/ && head -4 /home/node/.pi/agent/skills/compile-wiki/SKILL.md' </dev/null
-   ```
-
-   Expect `SKILL.md` non-empty with its frontmatter, and no `scripts/` directory.
-4. Confirm no `okf-lint` reference survives on the container side:
-
-   ```bash
-   grep -rn "okf-lint" pi/container/agent/ scripts/compile-wiki-container.sh
-   ```
-
-   Expect no matches.
-5. End-to-end on a cheap input rather than the 391 KB book in `md/`: put one
-   short Markdown file in a scratch folder and run
-   `./scripts/compile-wiki-container.sh <that-folder>`. In the transcript, check
-   specifically that:
-   - the agent issues a **`read` tool call on `SKILL.md`**, then one on
-     **`SPEC.md`** — no failed `bash`/`cat` attempt in between. A blocked-tool
-     error here is the signal that a `cat` fence survived somewhere;
-   - `okf/index.md` declares an `okf_version` matching the spec's §11
-     ("Versioning", `SPEC.md:391`) — proof the agent actually read the spec
-     rather than working from memory;
-   - new pages plus their `index.md` entries appear, with bundle-absolute links;
-   - `okf/log.md` exists, carries a `# Update Log` heading, and has **no `##`
-     headings and no dates**:
-
-     ```bash
-     grep -n '^## ' okf/log.md || echo "clean: no sub-headings"
-     ```
-
-   Run the driver a **second** time on a different short document and confirm the
-   new entries land **above** the previous ones, and that the earlier entries are
-   left untouched — that is the only real check of "newest first" and of
-   idempotency.
-
-   This writes into the real `okf/` (gitignored apart from `.okflintrc.json`), so
-   snapshot `okf/` first if you want the current wiki back.
-6. `make lint-okf` on the host — the independent check that replaces the
-   in-agent lint gate the container cannot run. Watch specifically for
-   **`log-date-format`** (severity `error`): any hit means the agent wrote a `##`
-   heading in `log.md`, which is the one way this design fails the linter.
-   `recommended-log` should now be silent, since `log.md` exists.
diff --git a/.claude/plans/observable-wiki-compilation.md b/.claude/plans/observable-wiki-compilation.md
deleted file mode 100644
index 623a2a5..0000000
--- a/.claude/plans/observable-wiki-compilation.md
+++ /dev/null
@@ -1,162 +0,0 @@
-# Observable `make wiki-sandbox`: raw event log + a Python trace viewer
-
-## Context
-
-A `make wiki-sandbox` run is a black box. Pi's print mode subscribes to the
-session event stream and **discards every event**, printing only the text blocks
-of the *final* assistant message ([print-mode.ts](https://github.com/earendil-works/pi/blob/main/packages/coding-agent/src/modes/print-mode.ts)) —
-so narration like "Now let me write chapters 7-10 and the glossary" is produced
-during the run but never reaches the terminal. That blindness is what made the
-stdin stall look identical to a working run, and it is why progress on a 15-minute
-compile can only be inferred from files appearing in `okf/`.
-
-`--verbose` does not help (the docs define it as forced verbose *startup*, i.e.
-the banner), and Pi deliberately ships no CLI-level tracing hook — its
-[observability design](https://github.com/earendil-works/pi/blob/main/packages/agent/docs/observability.md)
-exposes `subscribePiObservability()` to library embedders only, with OTel left to
-third-party wrappers. The supported route for a CLI run is `--mode json`, which
-streams every session event as JSON lines to stdout.
-
-Outcome: the driver writes the raw event stream to a log file, and a separate
-viewer renders it — live in a second terminal during a run, or after the fact
-against any saved log.
-
-This supersedes [.claude/plans/observe-pi-agent-sandbox.md](.claude/plans/observe-pi-agent-sandbox.md),
-an earlier unimplemented sketch of the same idea. Its jq filter is not used, but
-its findings are, and several of its "unverified" notes are now settled (below).
-
-## Decisions
-
-- The driver writes **raw JSONL only**; the compile terminal stays quiet apart
-  from a per-document heading naming the log file.
-- The viewer shows **narration + tool calls + per-turn usage**, plus compaction
-  and auto-retry warnings.
-- The viewer is **Python** (covered by `ruff check .` in `make lint`).
-- **Sandbox only** — `scripts/compile-wiki-container.sh` keeps its current
-  behaviour.
-
-## What is already verified
-
-- `--mode json` is dispatched at [main.ts:113](https://github.com/earendil-works/pi/blob/main/packages/coding-agent/src/main.ts#L113)
-  *before* the print check, so it is one-shot non-interactive on its own. Line
-  116 also shows Pi selects print mode automatically when stdout is not a TTY.
-- Each event is written with its own `write()`, so a redirect to a file grows in
-  real time — `--follow` works.
-- Real `usage` shape, read from a live session in the VM:
-  `{input, output, cacheRead, cacheWrite, reasoning, totalTokens, cost:{input, output, cacheRead, cacheWrite, total}}`.
-  Note `cost` is an **object** — the earlier plan's `\(.cost)` would have printed
-  `{...}`; use `cost.total`.
-- Assistant and toolResult messages carry a `timestamp` (epoch ms); bare events
-  such as `tool_execution_start` do not.
-- `jq` 1.7.1 is present at `/usr/bin/jq`, but is not used given the Python choice.
-
-Still unverified, and the reason the raw log is kept as ground truth: the exact
-shape of `tool_execution_end.result`. The session file only exposes toolResult
-*messages* (`{role, toolCallId, toolName, content:[{type, text}], isError, timestamp}`),
-which is probably but not certainly the same object. The viewer must use
-fallbacks, and the smoke test below settles it.
-
-## Changes
-
-### 1. `scripts/compile-wiki-sandbox.sh`
-
-Loop body becomes roughly:
-
-```bash
-log_dir="logs"
-mkdir -p "${log_dir}"
-stamp="$(date +%Y%m%d-%H%M%S)"
-
-for document in "${markdown_folder}"/*.md; do
-	slug="$(basename "${document}" .md)"
-	log="${log_dir}/${slug}-${stamp}.jsonl"
-	echo "Compiling document ${document}"
-	echo "  events → ${log}   (watch with: make wiki-watch)"
-	sbx exec "${kit_name}" -- pi \
-		--mode json \
-		--name "okf: ${slug}" \
-		-p "Load the compile-wiki skill: read ~/.pi/agent/skills/compile-wiki/SKILL.md, then follow it to compile ${document} into the OKF wiki under okf/." \
-		</dev/null >"${log}" \
-		|| { echo "Pi failed on ${document}; last events:" >&2; \
-		     python3 -u scripts/pi-trace.py --tail 20 "${log}" >&2; exit 1; }
-done
-```
-
-Points that matter:
-
-- `</dev/null` **stays** — it is what fixed the stdin stall.
-- `-p` is redundant under `--mode json` but harmless; keeping it means the script
-  still behaves if the mode flag is ever removed.
-- `--name "okf: <slug>"` labels the session so `sbx exec -it pi-kit -- pi -r`
-  finds it later in the picker.
-- Redirect, not `tee`: with "raw only" there is nothing to render inline, and a
-  plain `>` avoids `pipefail` interactions.
-
-### 2. `scripts/pi-trace.py` (new)
-
-`python3 -u scripts/pi-trace.py [FILE] [--follow] [--tail N]`
-
-- `FILE` defaults to the newest `logs/*.jsonl`; `-` reads stdin. `--follow` tails
-  a growing file, waiting up to ~60s for it to appear so it can be started before
-  the driver.
-- Standard library only. One `handle_<event_type>` mapping, unknown events
-  ignored so a Pi upgrade cannot crash the viewer.
-
-| Event | Rendered as |
-| --- | --- |
-| `session` | `── session <id8>  cwd=…` header |
-| `message_end` (assistant) | `💬 ` + text blocks, wrapped and indented |
-| `tool_execution_start` | `→ <tool>  <key arg>` — `command` / `path` / `file_path` / `pattern` / `query`, else compact JSON, clipped to ~100 chars |
-| `tool_execution_end` | `  ✓/✗ <tool>  <result summary>` from `.result.output // .result.content[].text // .result`, clipped to ~160 |
-| `turn_end` | `── turn  in=… out=… cost=$…` from `message.usage`, using `cost.total` |
-| `compaction_start` / `auto_retry_start` | `⚠ …` with reason / attempt |
-| `agent_end` | final summary: turns, tool calls, total tokens, total cost, elapsed |
-
-Timestamps: use `message.timestamp` when the event carries a message; otherwise
-wall clock in `--follow` mode, and carry forward the last known message timestamp
-when replaying a saved log. Narration is *not* aggressively clipped — it is the
-point of the exercise.
-
-### 3. `Makefile`
-
-Add a `wiki-watch` target (`python3 -u scripts/pi-trace.py --follow`), extend
-`.PHONY`, and mention it in the header comment block next to `wiki-sandbox`.
-
-### 4. `.gitignore`
-
-Add `logs/`.
-
-### 5. `README.md`
-
-Short subsection under the sandbox runtime: run `make wiki-sandbox` in one
-terminal and `make wiki-watch` in another; logs are per-document and disposable;
-`pi -r` inside the sandbox reopens a session in the full TUI for a rich
-post-mortem.
-
-## Verification
-
-1. **Smoke-test the event schema first** — this settles `tool_execution_end`
-   before the viewer is finished:
-
-   ```bash
-   sbx exec pi-kit -- pi --mode json --no-session \
-     "Run 'ls okf' with the bash tool, then reply DONE." </dev/null > /tmp/smoke.jsonl
-   echo "exit=$?"
-   jq -r '.type' /tmp/smoke.jsonl | sort | uniq -c
-   jq 'select(.type=="tool_execution_end") | .result' /tmp/smoke.jsonl | head -20
-   ```
-
-   Confirm the result keys, and whether a failed run exits non-zero in json mode
-   (the `||` branch in the driver depends on it; if json mode always exits 0,
-   detect failure by scanning the log for an error `stopReason` instead).
-
-2. `make lint` — ruff lints the new `scripts/pi-trace.py`, shellcheck the driver.
-   `make validate` — required by AGENTS.md for any change under `scripts/`.
-3. **Replay** an existing log end-to-end:
-   `python3 -u scripts/pi-trace.py logs/<file>.jsonl | head -40`.
-4. **Live**: `make wiki-sandbox` in one terminal, `make wiki-watch` in another.
-   Confirm narration appears within a second or two of the model emitting it (not
-   in a burst at the end), that the log grows, and that `✗` shows up for the
-   failed-`edit` case that occurs naturally on long chapters.
-5. **Interrupt path**: Ctrl-C a run and confirm the partial log still renders and
-   the viewer exits cleanly rather than tracebacking on a truncated last line.
diff --git a/.claude/plans/observe-pi-agent-sandbox.md b/.claude/plans/observe-pi-agent-sandbox.md
deleted file mode 100644
index f544e61..0000000
--- a/.claude/plans/observe-pi-agent-sandbox.md
+++ /dev/null
@@ -1,148 +0,0 @@
-# Making `make wiki-sandbox` observable
-
-Recommendation for seeing what the Pi agent actually does during a headless
-`sbx exec … pi -p …` run.
-
-## Why you currently see nothing
-
-Pi's print-mode source ([`print-mode.ts`](https://github.com/earendil-works/pi/blob/main/packages/coding-agent/src/modes/print-mode.ts))
-shows that with `-p`, Pi subscribes to the session event stream and **discards
-every event**, then prints only the text blocks of the final assistant message:
-
-```ts
-unsubscribe = session.subscribe((event) => {
-    if (mode === "json") { writeRawStdout(`${JSON.stringify(event)}\n`); }   // ← nothing happens in text mode
-});
-```
-
-So the blindness isn't `sbx`'s fault — it's the mode. `--verbose` won't help
-either; the docs define it as "force verbose *startup*" (banner/config echo),
-not a tool trace.
-
-## Options considered
-
-| Option | Verdict |
-| --- | --- |
-| `--mode json` — every event as JSON lines on stdout | **Best.** Live, structured, no VM changes |
-| `--mode rpc` — bidirectional protocol over stdin/stdout | Built for embedding Pi in a custom UI; overkill here |
-| Post-hoc session inspection (`pi -r`, `/export`, `pi --export s.jsonl out.html`) | Excellent *complement*, but not live |
-| A custom Pi extension that subscribes to events | Most flexible, most work — `--mode json` already gives the same event objects |
-| `sbx exec -it … pi` interactively | The manual route; doesn't fix `make wiki-sandbox` |
-
-## Recommendation: `--mode json`, formatted on the host, tee'd to a log
-
-`--mode json` is dispatched in [`main.ts:113`](https://github.com/earendil-works/pi/blob/main/packages/coding-agent/src/main.ts#L113)
-*before* the print check, so it is already one-shot non-interactive — `-p` is not
-needed alongside it. Each event is written with its own `write()` (a promise
-chain, no batching), so it streams in real time through the `sbx exec` pipe.
-
-You get `tool_execution_start` / `_end` (every bash command, every file write,
-and whether it errored), `message_end`, `turn_end` with token usage, plus
-`compaction_start` and `auto_retry_start` — exactly the events that silently eat
-a long run.
-
-Three edits:
-
-### 1. `scripts/pi-trace.jq` (new)
-
-Turns the event stream into a readable trace:
-
-```jq
-def clock: (now | localtime | strftime("%H:%M:%S"));
-def clip($n): tostring | gsub("\\s+"; " ") | if (length > $n) then .[0:$n] + "…" else . end;
-def tool_args: . as $a
-  | ($a.command // $a.path // $a.file_path // $a.pattern // $a.query // ($a | tojson)) | clip(100);
-def assistant_text: [.content[]? | select(.type == "text") | .text] | join("");
-
-if   .type == "session"              then "\(clock) ── session \(.id[0:8]) cwd=\(.cwd)"
-elif .type == "tool_execution_start" then "\(clock) → \(.toolName)  \(.args | tool_args)"
-elif .type == "tool_execution_end"   then "\(clock)   \(if .isError then "✗" else "✓" end) \(.toolName)  \((.result.output // .result.content // .result) | clip(160))"
-elif .type == "message_end" and .message.role == "assistant" then
-  (.message | assistant_text) as $t | if ($t|length) > 0 then "\(clock) 💬 \($t | clip(600))" else empty end
-elif .type == "turn_end"             then (.message.usage // {}) as $u
-  | "\(clock) ── turn end  in=\($u.input // "?") out=\($u.output // "?") cost=\($u.cost // "?")"
-elif .type == "compaction_start"     then "\(clock) ⚠ compacting context (\(.reason))"
-elif .type == "auto_retry_start"     then "\(clock) ⚠ retry \(.attempt)/\(.maxAttempts): \(.errorMessage | clip(120))"
-elif .type == "agent_end"            then "\(clock) ── agent end"
-else empty end
-```
-
-### 2. `scripts/compile-wiki-sandbox.sh` (lines 42–46)
-
-Replace the loop body:
-
-```bash
-log_dir="logs"
-mkdir -p "${log_dir}"
-
-for document in "${markdown_folder}"/*.md; do
-	slug="$(basename "${document}" .md)"
-	echo "Compiling document ${document}  (raw events: ${log_dir}/${slug}.jsonl)"
-	sbx exec "${kit_name}" -- pi \
-		--mode json \
-		--name "okf: ${slug}" \
-		"Read ${document} carefully. Then create or update the OKF wiki under okf/ following your instructions." \
-		| tee "${log_dir}/${slug}.jsonl" \
-		| jq -r --unbuffered -f scripts/pi-trace.jq
-done
-```
-
-`--unbuffered` matters — without it jq block-buffers into the terminal and you
-are back to staring at nothing. `pipefail` is already set, so a Pi failure still
-aborts the loop. `--name` labels the session so the picker below is navigable.
-
-### 3. `.gitignore`
-
-Add `logs/`.
-
-### Sample output
-
-Rendered against a synthetic stream built from the documented schema:
-
-```text
-17:28:19 ── session 3f2a9b10 cwd=/workspace
-17:28:19 → bash  cat SPEC.md
-17:28:19   ✓ bash  # OKF Specification Version: 1.4 lots and lots of text that goes on…
-17:28:19 → write  okf/index.md
-17:28:19   ✓ write  wrote 3 lines
-17:28:19   ✗ read  ENOENT: no such file
-17:28:19 💬 I read SPEC.md (v1.4) and updated okf/index.md accordingly.
-17:28:19 ── turn end  in=12045 out=842 cost=0.0123
-```
-
-The formatter was tested against that synthetic file, **not** a live run — the
-exact key names inside `result` and `usage` may differ slightly, which is why
-every field has a `//` fallback and the raw `.jsonl` is kept as ground truth.
-
-## Free complement: post-mortem in the TUI
-
-The script leaves the sandbox running detached, and sessions persist in
-`~/.pi/agent/sessions/` inside the VM. So after a bad run:
-
-```bash
-sbx exec -it pi-kit -- pi -r      # session picker → open the run in the full TUI
-                                  # then /tree to walk it, /export report.html
-```
-
-That gives the rich rendering (diffs, collapsible tool output) without re-running
-anything.
-
-## Two notes
-
-- `scripts/compile-wiki-container.sh` (lines 53–57) has the identical blind
-  spot. Per `AGENTS.md` the runtimes are deliberately independent, so it needs
-  the same edit applied by hand — the only difference is that `-xt bash` stays
-  and there is no `sbx exec` prefix.
-- Writing session files straight onto the host tree via
-  `--session-dir logs/sessions` was considered (the workspace is mounted rw, so
-  they would land in the working copy for `tail -f` and host-side `pi --export`).
-  Left out because how Pi resolves a *relative* `--session-dir` against the VM
-  cwd is unverified. Worth a one-off test if the HTML transcripts should land on
-  the host automatically.
-
-## Sources
-
-- [json.md](https://github.com/earendil-works/pi/blob/main/packages/coding-agent/docs/json.md)
-- [usage.md](https://github.com/earendil-works/pi/blob/main/packages/coding-agent/docs/usage.md)
-- [sessions.md](https://github.com/earendil-works/pi/blob/main/packages/coding-agent/docs/sessions.md)
-- [print-mode.ts](https://github.com/earendil-works/pi/blob/main/packages/coding-agent/src/modes/print-mode.ts)
diff --git a/.cursor/plans/google-style-guide-to-markdown.md b/.cursor/plans/google-style-guide-to-markdown.md
deleted file mode 100644
index bef274e..0000000
--- a/.cursor/plans/google-style-guide-to-markdown.md
+++ /dev/null
@@ -1,325 +0,0 @@
-# Google developer documentation style guide as one Markdown file
-
-## Context
-
-`md/` wants clean, structured Markdown. The Google developer documentation
-style guide (<https://developers.google.com/style>) is published as a DevSite
-book with no single-file download, so it needs a fetch-and-convert step of its
-own — a sibling to `pdf2md/`, which does the same job for PDFs. The output is
-one file in `md/`, and from there the existing pipeline takes over: `make
-wiki-sandbox` runs Pi once per `md/*.md` file and folds it into `okf/`.
-
-This supersedes
-[.cursor/plans/google_style_guide_scraper_plan.md](.cursor/plans/google_style_guide_scraper_plan.md).
-Its overall shape — discover, fetch, clean, convert, normalize links, assemble,
-validate — is right and is kept. Its facts about the site are not, and its
-reference script would not run. Both are corrected below against measurements
-taken from the live site on 2026-07-31.
-
-## What the site actually is (measured)
-
-Ten pages were fetched and analysed: the landing page, `word-list`, `tone`,
-`notices`, `lists`, `tables`, `code-samples`, `code-syntax`, `highlights`.
-
-- **70 pages, not "hundreds."** The book nav is server-rendered in the HTML of
-  every page: 71 `/style` anchors (one is the "Guides" tab, pointing at `/style`
-  a second time) inside `.devsite-book-nav-wrapper`, grouped by 10
-  `li.devsite-nav-heading` labels — Introduction, Key resources, General
-  principles, Language and grammar, Punctuation, Formatting and organization,
-  Linking, Computer interfaces, HTML and CSS, Names and naming. No JavaScript,
-  no lazy-loaded subtree: one fetch yields the whole ordered outline.
-- **`robots.txt` allows it, and nothing blocks a scraper.** `User-agent: *` with
-  a single `Disallow: /youtube/partner/`, plus a declared sitemap. Across ~15
-  requests with three different identities — a spoofed `Mozilla/5.0`, a truthful
-  `md2okf-web2md/0.1`, and no `User-Agent` header at all — every response was
-  `200`, including six pages back to back at 0.3 s intervals. No `429`, no
-  `Retry-After`, no challenge page. So there is no reason to spoof Mozilla as
-  the old plan did. **Untested:** a sustained run across all 70 pages, and
-  repeat runs from one IP. See Risks.
-- **Roughly 1.2 MB of article HTML.** Article bodies measured 5–33 KB each,
-  except `word-list` at 251 KB. Extrapolated: ~1.2 MB of body HTML, which
-  converts to about **0.5–0.9 MB of Markdown** — not the 1.5–4 MB the old plan
-  asserts as its size check.
-- **No shortcut exists.** `?print=1` returns the same single page (byte-for-byte
-  the same size), `/style/_book.yaml`, `/style/_toc.yaml`, `/llms.txt` and
-  `.txt` suffixes all 404, and `.md` / `?format=markdown` return the HTML page
-  with `Content-Type: text/html`. The `<devsite-llm-tools>` element on each page
-  is an empty JS-driven widget with no server endpoint behind it. Scraping HTML
-  is the only route.
-- **Every page carries an AI-generated summary inside the article body.**
-  `<devsite-key-takeaways-panel>` ("Page Summary", tooltip "Generated with AI")
-  sits inside `div.devsite-article-body` on all 10 sampled pages. It is not
-  source content and must be stripped — the old plan's removal list misses it,
-  along with `<devsite-recommendations>`, `<devsite-feedback>` and
-  `<devsite-thumb-rating>`, which are also inside or adjacent to the body.
-- **90% of internal links carry a fragment.** 838 of 936 `/style` links in the
-  sampled bodies point at `#anchor` targets. The old plan strips fragments and
-  maps every link to a bare page slug, which throws away almost all of the
-  cross-referencing the guide is built on.
-- **Anchor ids collide across pages.** `examples` occurs on 3 sampled pages;
-  `politeness-and-use-of-please`, `list-or-table`,
-  `some-things-to-avoid-where-possible` and
-  `some-techniques-and-approaches-to-consider` each on 2. Merging 70 pages into
-  one file therefore needs namespaced anchors, not the raw ids.
-- **The word list is a definition list, and it is 40% of the guide.** 26 `<dl>`
-  blocks, 598 `<dt>` terms, every one with an `id` (only 2 are not slug-safe:
-  `+` and `google-play-services-SDK`). markdownify 1.2.3 does convert `dt`/`dd`,
-  but into Pandoc/PHP-Markdown-Extra syntax (`term` then `:   definition`,
-  source lines 521–556) that GitHub and most renderers do not support.
-- **Some meaning exists only in CSS.** Counted in the sample:
-  `<span class="icon-dontuse">` 148 times, `icon-avoid` 93, plus `icon-android`
-  27, `icon-cloud` 9, `icon-workspace` 3. These spans are **empty** — a naive
-  conversion silently drops "don't use this term" from hundreds of word-list
-  entries. By contrast `compare-better` (158) and `compare-worse` (91) wrap the
-  words "Recommended" / "Not recommended", so those survive on their own.
-- **`material-icons` leaks the word "link" 600+ times.** Each `<dt>` ends with
-  `<a href="#id"><span class="material-icons" aria-hidden="true">link</span></a>`;
-  `word-list` alone has 603 occurrences.
-- **Code blocks have no language.** Every one is
-  `<devsite-code><pre class="devsite-click-to-copy">…</pre></devsite-code>`, and
-  no `lang-*` or `language-*` class appears anywhere in the sample. The old
-  plan's language extraction has nothing to extract; plain fences are correct.
-- **Body headings run h2 to h4 only** (60 h2, 37 h3, 5 h4 in the sample), and
-  there are 10 images, all with site-relative `src`.
-
-## Where the old reference script breaks
-
-- `convert_aside(self, el, text, convert_as_inline)` is the pre-1.0 markdownify
-  signature. Since 1.0 the hook is `convert_<tag>(self, el, text, parent_tags)`
-  (markdownify is at 1.2.3), so the method would never be called with the
-  arguments it expects.
-- Several string literals contain raw newlines inside non-triple quotes, and an
-  f-string embeds unescaped `"` — the file as written is not valid Python.
-- It declares async, concurrency, caching and retries in the prose, then fetches
-  with a bare synchronous `httpx.get` in a loop, with no cache and no retry.
-- `nav = soup.find("ul", class_="devsite-nav-section-list") or soup` — that class
-  does not exist on the page, so the fallback silently harvests every `/style`
-  link in the whole document, in DOM order rather than nav order, and loses the
-  10 section groups entirely.
-- Nothing removes the AI summary panel, and nothing handles `dl`/`dt`/`dd`, the
-  empty semantic spans, or the icon text.
-
-## `marker` and LLMs: no for converting, yes for checking
-
-`marker` is the wrong tool here, and the reason is structural, not a matter of
-tuning. Its job is to **recover** structure a PDF has already lost — reading
-order, table geometry, headings inferred from font size — and it leans on a
-language model to guess well. This source has lost nothing: DevSite ships
-semantic HTML carrying the exact `id` attributes that 838 cross-links depend on.
-Routing HTML through a headless-browser PDF print and then `marker` would
-destroy every `href` and anchor, turn a 598-term definition list into guessed
-headings or tables, cost one LLM call per page, and produce a different file on
-every run — so `git diff` on a refresh would be unreadable. Deterministic HTML
-parsing is lossless, free and fast (~25 s for 70 pages).
-
-What `pdf2md` does contribute is its **checking** chain, which applies verbatim:
-`prettier --check`, `markdownlint-cli2` and `cspell` over the generated file, as
-documented in [pdf2md/README.md](pdf2md/README.md). Reuse it rather than
-inventing a new "syntax check."
-
-An LLM is still worth one optional, off-by-default pass: `--audit N` samples N
-pages, hands the model the cleaned HTML text beside the emitted Markdown, and
-asks what content disappeared. That is exactly the class of bug the CSS-only
-`icon-dontuse` spans represent, and it is cheap on a sample. It must stay out of
-the write path so the output stays reproducible.
-
-The genuinely useful LLM stage is the one that already exists downstream:
-`make wiki-sandbox` folds the file into OKF. Keeping the scraper deterministic
-leaves that as the only nondeterministic stage in the chain. One caveat worth
-watching: a ~0.6 MB source is roughly 150K tokens for a single Pi run — fine on
-the 1M-context catalogue models the README describes, but the first run is the
-place to confirm it.
-
-## Decisions
-
-- **One output file**, `md/GoogleStyleGuide.md`, matching
-  the existing `md/TheEconomistStyleGuide2023.md` naming.
-- **New `web2md/` directory**, mirroring `pdf2md/`: one Python module plus a
-  README, wired into `make lint` and a `make` target.
-- **Sequential fetch with an on-disk cache.** 70 pages at ~200 ms is ~25 s; the
-  old plan's concurrency is complexity without a payoff, and a cache makes
-  re-runs of the conversion free.
-- **Deterministic output.** Same input, same bytes, so a refresh diffs cleanly.
-- **Scraper dependencies in their own dependency group**, so CI's
-  `uv run --only-group dev ruff` stays untouched and nobody needs `marker-pdf`
-  (and torch) to scrape a website.
-
-## Changes
-
-### `web2md/web2md.py` (new)
-
-One module, argparse CLI, six stages. DevSite-specific selectors live in
-constants at the top so pointing it at another DevSite book is a config change.
-
-```python
-BASE = "https://developers.google.com"
-BOOK = f"{BASE}/style"
-LOCALE = {"hl": "en"}  # pin the locale; the page honours ?hl=en
-UA = "md2okf-web2md/0.1 (+https://github.com/<owner>/md2okf)"
-NAV = ".devsite-book-nav-wrapper"
-BODY = "div.devsite-article-body"
-DROP = (
-    "devsite-key-takeaways-panel",
-    "devsite-recommendations",
-    "devsite-feedback",
-    "devsite-thumb-rating",
-    "devsite-toc",
-    ".devsite-rating-container",
-    ".devsite-content-footer",
-    ".material-icons",
-    "[aria-hidden=true]",
-    "script",
-    "style",
-    "noscript",
-    "iframe",
-)
-```
-
-1. **Discover.** Fetch `BOOK`, walk `li` elements under `NAV` in document
-   order; `li.devsite-nav-heading` sets the current section, every other
-   `li.devsite-nav-item` appends `{section, title, slug, url}`. Dedupe by URL,
-   first-seen order. Assert `60 <= len(pages) <= 200` and `len(sections) >= 5`,
-   and fail loudly otherwise — a silent 3-page run is the worst outcome.
-2. **Fetch.** One `httpx.Client(headers={"User-Agent": UA},
-   follow_redirects=True)`, `params=LOCALE`, ~200 ms between requests. Cache to
-   `web2md/cache/{slug}.html`; only `--refresh` re-fetches. Retry 429 and 5xx
-   with exponential backoff in an explicit loop — note that
-   `httpx.HTTPTransport(retries=…)` retries connection errors only, not status
-   codes. Two rules make throttling impossible to mistake for success, since a
-   challenge or error page still parses as HTML and would otherwise yield a
-   hollow section: **cache only `200` responses**, so a throttled reply never
-   becomes a poisoned cache entry, and **treat a missing `BODY` as a hard
-   error** that stops the run, naming the page.
-3. **Clean.** Take `BODY`, drop everything in `DROP`, unwrap `<devsite-code>` to
-   the `<pre>` it contains. Before dropping, replace the empty semantic spans
-   with text: `icon-dontuse` → `Don't use:`, `icon-avoid` → `Avoid:`,
-   `icon-android`/`icon-cloud`/`icon-workspace` → `(Android)`/`(Cloud)`/
-   `(Workspace)`. Take the page title from the nav, not from the `h1` — the `h1`
-   lives outside the article body.
-4. **Convert.** Subclass `MarkdownConverter` (1.2.3 signature,
-   `convert_<tag>(self, el, text, parent_tags)`; hyphens in tag names become
-   underscores in method names, so `<devsite-code>` would be
-   `convert_devsite_code` if it were not unwrapped in stage 3).
-   - `convert_aside`: `note` → `> [!NOTE]`, `caution` → `> [!CAUTION]`,
-     `warning` → `> [!WARNING]`, `success` → `> [!TIP]`, anything else →
-     `> [!NOTE]` plus a warning to stderr. All four classes are confirmed
-     present on `/style/notices`.
-   - `convert_dt` / `convert_dd`: override the Pandoc-style default. Emit
-     `<a id="{anchor}"></a>**{term}**` followed by the definition as an indented
-     paragraph, so the 598 word-list ids stay reachable and the file renders on
-     GitHub.
-   - `convert_hN`: shift by +2 and **reuse DevSite's own `id`**, never a
-     synthesized slug — the 838 fragment links target those exact strings,
-     including oddities like `formatting,-punctuation,-and-organization`.
-   - `convert_img`: absolutise `src` against `BASE`, keep the alt text.
-   - `<pre>`: plain triple-backtick fences, no language tag.
-5. **Anchors and links.** Collect every `id` in every cleaned body (759 in the
-   sample — sections, headings and `dt` terms all carry them) into one map:
-   `(slug, id) -> f"{slug}--{sanitize(id)}"`, where `sanitize` maps anything
-   outside `[A-Za-z0-9._-]` to a stable replacement (2 cases in the whole guide:
-   `+` and `google-play-services-SDK`). Emit an explicit `<a id="…"></a>` at
-   each target, then rewrite:
-   - `/style/x#y` and `https://developers.google.com/style/x#y` → `#x--y`
-   - `/style/x` → `#x`
-   - bare `#y` on page `x` → `#x--y`
-   - anything else, or a `/style` target missing from the map → absolute URL,
-     logged to stderr with a count.
-6. **Assemble.** Frontmatter written directly (five static lines: `title`,
-   `source`, `generated_at`, `pages`, `format` — no `pyyaml` dependency for
-   that), then the `h1` document title, then a TOC grouped by the 10 nav
-   sections, then per page an `h2` section heading, an `h3` page title, a
-   `*Source: <url>*` line and the
-   body. Depth budget: `#` document, `##` section, `###` page, body h2→h4,
-   h3→h5, h4→h6. That lands exactly on markdownify's h6 clamp, which is why
-   validation asserts no body heading is deeper than h4.
-
-### `web2md/README.md` (new)
-
-Same shape as `pdf2md/README.md`: what it does, the commands, and the
-`prettier` / `markdownlint-cli2` / `cspell` check chain pointed at the generated
-file. State plainly that the output is a snapshot of a living document and that
-`--refresh` is how you update it.
-
-### `pyproject.toml`
-
-```toml
-[dependency-groups]
-dev = ["ruff>=0.8"]
-web2md = [
-  "httpx>=0.28.1",
-  "beautifulsoup4>=4.15.0",
-  "lxml>=6.1.1",
-  "markdownify>=1.2.3",
-]
-```
-
-Run it as `uv run --group web2md python web2md/web2md.py`. `pyyaml`
-from the old plan is not needed.
-
-### `Makefile`
-
-- Add `"web2md/README.md"` to the `markdownlint-cli2` list in `lint`.
-- Add a target:
-
-```make
-# Fetch the Google developer documentation style guide into md/ as one file.
-style-guide:
-	uv run --group web2md python web2md/web2md.py
-```
-
-`ruff check .` already covers the new module.
-
-### `.gitignore`
-
-Add `web2md/cache/` — raw HTML snapshots are throwaway. The generated Markdown
-needs no rule; it lands in `md/`, which is ignored already.
-
-### `.github/workflows/ci.yml` (optional)
-
-The `paths` anchor lists no `pdf2md/**` today, so a `web2md/**` entry is a
-choice rather than a fix. Add it if you want CI to lint scraper-only changes.
-
-## Validation
-
-In-script, exiting non-zero on failure:
-
-1. Pages written equals pages discovered; list any that failed.
-2. Every `#…` link target resolves to an emitted anchor; print the offenders.
-3. No `devsite-`, no `material-icons`, and no stray Pandoc definition lines (a
-   colon followed by indent) survive in the output; no body heading deeper than
-   `h4` was seen.
-4. Size between 0.4 MB and 1.2 MB, from the ~1.2 MB of measured body HTML —
-   replacing the old 1.5–4 MB guess.
-5. Word-list term count within a tolerance of the 598 measured, since a silent
-   `dl` regression there would gut 40% of the guide.
-
-Then, by hand on the host, the `pdf2md` chain:
-
-```bash
-prettier --check md/GoogleStyleGuide.md
-markdownlint-cli2 md/GoogleStyleGuide.md
-cspell md/GoogleStyleGuide.md
-```
-
-Finish with `make lint`, and a `make wiki-sandbox` run to confirm the file works
-as a Pi source document. `make validate` is not needed: nothing under `pi/` or
-`scripts/` changes.
-
-## Risks
-
-- **The nav is one selector deep.** Everything hinges on
-  `.devsite-book-nav-wrapper`. Hence the count assertions in stage 1: a DevSite
-  redesign should stop the run, not quietly produce a stub.
-- **Google could throttle a run, even though nothing suggests it will.** The
-  15-request probe above was clean at three different identities, but a full
-  70-page pass has not been measured, and neither has a second pass from the
-  same IP. The stage 2 rules are the whole defence: back off on `429`, cache
-  only `200`s, and stop on a missing article body. If a run ever does trip a
-  limit, the first move is a longer delay, not a browser or a proxy — 70 pages
-  is small enough that politeness costs nothing.
-- **The guide changes under you.** The output is a dated snapshot. `whats-new`
-  is part of the book, so the file records its own currency.
-- **CSS-only semantics may be wider than measured.** The 10-page sample found
-  five such classes. The optional `--audit` pass exists for exactly this, and a
-  cheap guard is to log every `class` attribute value dropped during cleaning,
-  so a new one shows up as a line of output rather than as missing content.
diff --git a/.cursor/plans/google_style_guide_scraper_plan.md b/.cursor/plans/google_style_guide_scraper_plan.md
deleted file mode 100644
index 67e7214..0000000
--- a/.cursor/plans/google_style_guide_scraper_plan.md
+++ /dev/null
@@ -1,313 +0,0 @@
-# Coding Agent Specification: Google Style Guide Single-File Markdown Generator
-
-## 1. Problem Statement
-
-### 1.1 Overview
-The **Google Developer Documentation Style Guide** (hosted at `https://developers.google.com/style`) is an essential reference for technical writers, software engineers, and AI prompt engineers. However, Google publishes this guide across **hundreds of discrete web pages** without providing an official single-file download (such as a consolidated PDF or Markdown document).
-
-### 1.2 Target Audience & Use Case
-A single-file Markdown (`.md`) document is required to:
-1. Provide **offline accessibility** and fast local text searching.
-2. Serve as a **clean context document** for Large Language Models (LLMs) and RAG (Retrieval-Augmented Generation) pipelines.
-3. Enable easy version-controlled local archival and custom printing.
-
-### 1.3 Technical Challenges
-- **Navigation Hierarchy:** Page order matters. The scraper must parse the left-hand navigation sidebar to retain the intended reading sequence.
-- **Boilerplate & Noise Removal:** Google DevSite pages contain extensive non-content UI (navigation bars, search boxes, header/footer, rating widgets, sidebars, cookie banners).
-- **Custom DOM Elements:** DevSite uses custom tags and CSS classes for callouts (`aside class="note"`, `aside class="caution"`), custom code tabs, and structured tables.
-- **Internal Link Disruption:** Relative links pointing to `https://developers.google.com/style/...` will break unless converted into local section anchors (`#...`).
-- **Rate Limiting & Politeness:** Fetching hundreds of pages sequentially requires request throttling, proper User-Agent headers, and retry logic.
-
----
-
-## 2. Solutions Overview
-
-### Option Comparison
-
-| Approach | Pros | Cons | Decision |
-|---|---|---|---|
-| **A. Static HTML Scraper (`httpx` + `BeautifulSoup` + `markdownify`)** | Fast, lightweight, pure Python, highly customizable AST conversion. | Requires manually mapping custom DOM elements. | **RECOMMENDED** |
-| **B. Headless Browser (`Playwright` + Node/Python + Turndown)** | Renders JavaScript-heavy elements automatically. | Slower, heavier resource usage, unnecessary overhead for static text pages. | Secondary fallback |
-| **C. Shell Pipeline (`wget` + `pandoc`)** | Simple one-liner execution. | Messy output, preserves UI clutter, mangles navigation order, poor callout handling. | Rejected |
-
-### Recommended Solution Architecture
-A modular Python script operating in 6 sequential stages:
-1. **Discover:** Extract page hierarchy from sidebar navigation.
-2. **Fetch:** Retriable HTTP fetches with concurrent batching and caching.
-3. **Clean:** Target main content containers and strip site chrome/UI.
-4. **Transform:** Convert HTML AST to clean Markdown with custom rules for code and callouts.
-5. **Normalize:** Update internal cross-references to point to document anchors.
-6. **Compile:** Assemble master TOC, frontmatter, and output single `.md` file.
-
----
-
-## 3. Detailed Staged Implementation Plan
-
-```
-[ Stage 1: Setup ] ──> [ Stage 2: Crawl Nav Tree ] ──> [ Stage 3: Fetch & Clean HTML ]
-                                                                     │
-[ Stage 6: Validation ] <── [ Stage 5: Compile & Normalize ] <── [ Stage 4: HTML -> MD ]
-```
-
-### Stage 1: Environment & Dependency Setup
-Set up a python environment with required libraries:
-- `httpx` (Async HTTP client with HTTP/2 and retry capabilities)
-- `beautifulsoup4` + `lxml` (Fast HTML parsing and tree navigation)
-- `markdownify` (HTML to Markdown converter with custom subclassing)
-- `pyyaml` (Frontmatter generation)
-
-#### Required Directory Layout
-```text
-scraper/
-├── cache/                  # Downloaded raw HTML files (prevents refetching)
-├── output/                 # Output directory for final MD file
-├── config.py               # CSS selectors, user agents, rate limits
-├── parser.py               # Custom HTML-to-MD rules
-└── main.py                 # CLI entrypoint and orchestrator
-```
-
----
-
-### Stage 2: Sitemap & Hierarchy Discovery
-**Objective:** Parse the left navigation menu on `https://developers.google.com/style` to build an ordered list of URLs with section levels.
-
-#### Step-by-Step Instructions:
-1. Fetch the main index page `https://developers.google.com/style`.
-2. Locate the navigation tree element:
-   - Primary selector: `ul.devsite-nav-section-list` or `nav.devsite-section-nav`.
-3. Extract all `<a>` tags with `href` attributes starting with `/style`.
-4. Store URLs sequentially in a list of structured dicts:
-   ```python
-   page_item = {
-       "title": "Voice and tone",
-       "url": "https://developers.google.com/style/voice",
-       "level": 2,  # Depth in nav hierarchy
-       "slug": "voice"
-   }
-   ```
-5. Deduplicate links while strictly maintaining first-seen order.
-
----
-
-### Stage 3: Fetching & DOM Cleaning
-**Objective:** Fetch raw HTML for each page, store in cache, and strip non-article elements.
-
-#### Step-by-Step Instructions:
-1. **Caching:** Check if `cache/{slug}.html` exists. If not, fetch via HTTP request with `User-Agent: Mozilla/5.0 ...`.
-2. **Rate Limiting:** Implement a 200–500ms delay between requests to avoid 429 throttling.
-3. **Locate Core Content:**
-   - Target container: `article.devsite-article` or `div.devsite-article-body`.
-4. **De-cluttering (Removal List):**
-   Remove the following selectors before processing:
-   - `nav`, `header`, `footer`
-   - `.devsite-rating-container` (Feedback widgets)
-   - `.devsite-toc` (In-page right-hand TOC)
-   - `.devsite-content-footer`
-   - `script`, `style`, `noscript`
-   - Buttons like "Copy code" or print triggers.
-
----
-
-### Stage 4: Custom HTML-to-Markdown Conversion
-**Objective:** Transform cleaned HTML DOM into clean Markdown while handling Google-specific elements.
-
-#### Custom Tag Handling Rules:
-
-1. **Headings (`<h1>` to `<h6>`):**
-   - Shift header ranks down by 1 level if using `# Title` for the main page document.
-   - Inject HTML anchor IDs or explicit slug targets: `<a id="slug-heading"></a>`.
-
-2. **Callout Boxes / Notices (`<aside>`):**
-   Google uses `<aside class="note">`, `<aside class="caution">`, `<aside class="key-point">`.
-   - Convert to GitHub-Flavored Markdown (GFM) admonition syntax:
-     ```markdown
-     > [!NOTE]
-     > Note text content here...
-     ```
-
-3. **Code Blocks (`<pre>`, `<code>`):**
-   - Extract language class if present (e.g. `class="lang-py"` -> `py`).
-   - Wrap in triple backticks ` ```python ... ``` `.
-
-4. **Tables (`<table>`):**
-   - Convert standard `<table>` / `<tr>` / `<th>` / `<td>` to Markdown pipe tables.
-   - Clean up interior multiline HTML inside table cells to single-line or `<br>`.
-
----
-
-### Stage 5: Link Normalization & Single-File Compilation
-**Objective:** Assemble all parsed sections into a single monolithic `.md` file and fix internal hyperlinks.
-
-#### Step-by-Step Instructions:
-1. **Frontmatter:** Inject YAML frontmatter at top of document:
-   ```yaml
-   ---
-   title: Google Developer Documentation Style Guide
-   source: https://developers.google.com/style
-   generated_at: YYYY-MM-DD
-   format: Single-file Markdown
-   ---
-   ```
-2. **Master Table of Contents:**
-   Generate a master TOC linking to every page heading using synthesized anchors (e.g. `#voice-and-tone`).
-3. **Link Rewriting Rule:**
-   - External links (`https://...` outside `/style`) -> Keep as-is.
-   - Internal links (`/style/voice` or `https://developers.google.com/style/voice`) -> Convert to local Markdown anchor links `[Voice and tone](#voice)`.
-4. **Page Dividers:**
-   Insert page section headers and horizontal rules between scraped pages:
-   ```markdown
-   ---
-   # Page Title
-   *Source: https://developers.google.com/style/page*
-   
-   [Page Content]
-   ```
-
----
-
-### Stage 6: Validation & Verification
-**Objective:** Guarantee the quality and completeness of the generated file.
-
-#### Automated Checks to Implement in Script:
-1. **Missing Content Check:** Ensure page count in output matches total discovered URLs.
-2. **Unbroken Links Check:** Verify all internal `#anchor` links point to valid targets within the generated document.
-3. **File Size Check:** Expected size is typically between 1.5 MB and 4.0 MB.
-4. **Syntax Check:** Ensure code blocks and tables do not break Markdown rendering syntax.
-
----
-
-## 4. Ready-to-Run Reference Implementation Script
-
-Below is a complete, runnable Python script that implements the above specification. A coding agent can execute or extend this directly:
-
-```python
-import os
-import re
-import time
-from urllib.parse import urljoin, urlparse
-import httpx
-from bs4 import BeautifulSoup
-from markdownify import MarkdownConverter
-
-BASE_URL = "https://developers.google.com/style"
-OUTPUT_FILE = "google_developer_style_guide.md"
-HEADERS = {"User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64)"}
-
-class GoogleStyleGuideConverter(MarkdownConverter):
-    def convert_aside(self, el, text, convert_as_inline):
-        classes = el.get("class", [])
-        kind = "NOTE"
-        if "caution" in classes or "warning" in classes:
-            kind = "WARNING"
-        elif "key-point" in classes:
-            kind = "IMPORTANT"
-        
-        lines = text.strip().split("
-")
-        quoted = "
-> ".join(lines)
-        return f"
-
-> [!{kind}]
-> {quoted}
-
-"
-
-def get_navigation_urls():
-    print("Fetching navigation index...")
-    res = httpx.get(BASE_URL, headers=HEADERS, follow_redirects=True)
-    soup = BeautifulSoup(res.text, "lxml")
-    
-    urls = []
-    # Find navigation items
-    nav = soup.find("ul", class_="devsite-nav-section-list") or soup
-    for a in nav.find_all("a", href=True):
-        href = a["href"]
-        if href.startswith("/style") or href.startswith(BASE_URL):
-            full_url = urljoin(BASE_URL, href).split("#")[0]
-            if full_url not in urls:
-                urls.append(full_url)
-    print(f"Discovered {len(urls)} pages in navigation.")
-    return urls
-
-def clean_and_convert_html(html_content, page_url):
-    soup = BeautifulSoup(html_content, "lxml")
-    
-    # Extract main article
-    article = soup.find("article", class_="devsite-article") or soup.find("div", class_="devsite-article-body")
-    if not article:
-        article = soup.find("main") or soup.body
-
-    # Strip junk elements
-    for selector in [
-        "nav", ".devsite-rating-container", ".devsite-toc",
-        ".devsite-content-footer", "script", "style", "iframe"
-    ]:
-        for el in article.select(selector):
-            el.decompose()
-
-    # Rewrite internal links to local anchors
-    for a in article.find_all("a", href=True):
-        href = a["href"]
-        if href.startswith("/style") or href.startswith(BASE_URL):
-            parsed = urlparse(href)
-            slug = parsed.path.strip("/").replace("/", "-")
-            a["href"] = f"#{slug}" if slug else "#introduction"
-
-    converter = GoogleStyleGuideConverter(heading_style="ATX")
-    return converter.convert(str(article))
-
-def build_single_file_md():
-    urls = get_navigation_urls()
-    if not urls:
-        urls = [BASE_URL] # Fallback
-        
-    md_sections = []
-    md_sections.append("---
-title: Google Developer Documentation Style Guide
-source: https://developers.google.com/style
----
-
-")
-    md_sections.append("# Google Developer Documentation Style Guide
-
-")
-
-    for idx, url in enumerate(urls, start=1):
-        parsed = urlparse(url)
-        slug = parsed.path.strip("/").replace("/", "-") or "introduction"
-        print(f"[{idx}/{len(urls)}] Processing: {url}")
-        
-        try:
-            res = httpx.get(url, headers=HEADERS, follow_redirects=True)
-            if res.status_code == 200:
-                md_content = clean_and_convert_html(res.text, url)
-                md_sections.append(f"
-
-<a id="{slug}"></a>
-
----
-
-")
-                md_sections.append(md_content)
-            time.sleep(0.2)
-        except Exception as e:
-            print(f"Error fetching {url}: {e}")
-
-    with open(OUTPUT_FILE, "w", encoding="utf-8") as f:
-        f.writelines(md_sections)
-    print(f"Successfully generated single-file style guide: {OUTPUT_FILE}")
-
-if __name__ == "__main__":
-    build_single_file_md()
-```
-
----
-
-## 5. Summary Checklist for Agent Execution
-
-- [ ] Execute script to discover navigation links.
-- [ ] Verify HTML sanitization rules strip header, footer, and navigation.
-- [ ] Confirm custom `<aside>` conversion produces GFM callout blockquotes.
-- [ ] Ensure internal links are mapped to `#slug` local anchors.
-- [ ] Validate final single `.md` file size and content integrity.