From f596d18733cbc1480c392085d21bd02ddbfc2d1b Mon Sep 17 00:00:00 2001 From: BY571 Date: Wed, 19 Aug 2026 18:01:06 +0200 Subject: [PATCH] drop the em dashes from the README Seven of them, all doing work a comma, a colon or a full stop does better. The 73 hyphens stay: they are inside flags (`--json`), node ids (`hyp-self-consistency`) and compound words, where they are part of the token rather than punctuation. One rewrite broke the graph.yaml example and the doc tests caught it: replacing an em dash with a colon inside an unquoted YAML value is a second mapping separator, so `question: what this graph exists to answer: a question...` no longer parses. That is the whole argument for keeping the README's examples executable, demonstrated on a one-character edit. --- README.md | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/README.md b/README.md index ad7bed0..07d504d 100644 --- a/README.md +++ b/README.md @@ -12,7 +12,8 @@ back, and the script that killed it. Nothing is deleted and nothing is overwritt correction is a new node that supersedes the old one, so six months later the graph can tell you not just what you believe but what you already ruled out and why. -It exists because research loops — human or agent — forget. They re-propose an idea that +It exists because research loops forget, whether they are run by a human or an agent. +They re-propose an idea that was settled last month under a different name, and they file the wins while the failures evaporate. knoten makes the failure the artifact: a claim cannot be marked alive unless it cites a test it survived, and a dead end has to say what would reopen it. Your graph @@ -87,7 +88,7 @@ knoten hook # make `git commit` refuse a broken graph knoten viz --open # the whole graph as one HTML file ``` -Every read command takes `--json`. Exit `0` succeeded, `1` refused — and a refusal is the +Every read command takes `--json`. Exit `0` succeeded, `1` refused, and a refusal is the feature: read it, fix the node, run it again. ## Rules are data @@ -108,7 +109,7 @@ rules: ``` The first is the safety mechanism: a good-looking result that was never checked cannot -quietly become a finding. The second is what makes a dead end *reusable* — once the cause +quietly become a finding. The second is what makes a dead end *reusable*. Once the cause is a field rather than a sentence, the question you ask six months later is a query: ```bash @@ -120,8 +121,8 @@ than new types: ```yaml node_types: - question: what this graph exists to answer — a question, a statement or a task - source: where the work came from — a paper, dataset, search, or your own intuition + question: what this graph exists to answer, be it a question, statement or task + source: where the work came from, such as a paper, dataset or your own intuition hypothesis: a falsifiable claim derived from an idea gate: a standing rule every claim must survive; a bar, not a stage statuses: [open, alive, dead, retracted, superseded, active] @@ -130,7 +131,7 @@ tags: [decoding, reasoning, prompting, evaluation] ## For agents -[`SKILL.md`](SKILL.md) is how a coding agent learns knoten — point Claude Code, or +[`SKILL.md`](SKILL.md) is how a coding agent learns knoten. Point Claude Code, or anything with a shell, at it. It teaches the loop above, which types of node a graph holds and which way an edge points.