Repository navigation
2026-usrse: con-duct poster for the US-RSE'26 poster session - #34
Merged
Merged
Conversation
HTML mockup of the US-RSE'26 con-duct poster, sized for a 42 x 36 in fabric print: @page sets the page size and a print stylesheet zooms the 2158 x 1850 px layout onto it, so headless chromium --print-to-pdf gives one page. Title row: author block, title with the con-duct logo inline, then the Dartmouth D-Pine, CON and ReproNim logos and QR placeholders. Assets include the con-duct overhead and fMRIPrep plots. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
Two QR codes replace the placeholders: the con/duct repository and the abstract's Zenodo DOI (10.5281/zenodo.22833931). Generated as SVG with `segno --scale 10 --border 0 --error H <url>`; both decode to their URL from the printed PDF. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
The D-Pine joins the third-party logos marker. The con-duct logo is a bitmap trace of an image of unrecorded origin (con/duct#207), so it gets the NOASSERTION mixed marker rather than the catch-all. The ReproNim logo is the already-annotated pics/borrowed/repronim_logo.svg, and the CON logos stay under the catch-all, being our own. `reuse lint` passes. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
Middle column: what con-duct is (capture, search, plot). Left: Try it with a real hello-world run, You drive / the agent drives with simulated chats, and the plot. Right: DataLad with a real datalad run commit, adopted inside tools (BABS, ReproNim/containers, ReproStim), comparison table. Footer: overhead and limitations. The k8s case study is cut and nebari is gone from the poster. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
- Claude Design graphic for the middle column (capture, then what the human or agent asks of the records), pasted as static markup and scaled by script; it replaces the HTML capture/search prose. - Claude Design header: teal band, duct logo tile, QR tiles. - New "Search is a Python expression" box with real con-duct ls output over the con/duct-gallery records. - Plot moves to the left column; Try it and the chat are trimmed; the Try it and DataLad runs are shown dated 2026-09-30. - Limitations trimmed and reworded around ps CPU statistics. - Logos move from the header to the footer. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
…message Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
… cards - The agent-drives panel shows real runs and a con-duct ls query that runs as written (paths before -F). - Plot swapped to the con/duct-gallery ds002785 ps-cpu-timepoint SVG. - Search questions as tinted cards; caption cut. - Overhead caption drops the records path; table right-aligned. - Limitations: macOS sampling overhead. - Comparison table: the partial marker is an inline SVG (the U+25D0 glyph fell back to a tiny sliver). Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
duct deletes the logs of a command that fails in under 3 s by default, so a fast-failing reproducer would leave nothing to list. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
Without the extras, con-duct ls prints a plain list (no pyout) and con-duct plot fails for lack of matplotlib. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
The sampler supports only Linux and Darwin (_SUPPORTED_SYSTEMS). Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
The old wording (do not wrap such commands) contradicted the abstract, which says output can be skipped for sensitive commands. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
GNU time reports wall, user and sys time plus max RSS, not one number, and tee is not a one-word prefix. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
The header and the graphic already use Plex; the body fell back to Noto. Inline code no longer breaks at the hyphen in con-duct. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
A one-line headline (about 2-3 % of one core, 37 MB, 1 to 256 processes), a plain table header, and a shorter method caption that keeps the 3.3 % whole-run figure and the slow-rise hedge. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
It printed at about 13 pt. The tool examples under the column headers stay. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
- Every font size raised to at least 18 pt at print (13 px in the poster, 24 and 25 design px in the scaled graphic and header). - Graphic: taller file boxes and wider question cards so the larger text fits; the canvas copy of the graphic is not updated. - Try it runs plain duct sleep 2 (real run, shown dated 2026-09-30). - Agent-drives commands use --f-t 0, the short form. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
The What is con-duct? paragraph goes from 14 to 17 px (about 24 pt at print); the graphic below moves up into its own top margin so the middle column still fits. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
Drop the sort example; the other three queries stack full width with each command on one line. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
No -m, and the command writes just hello. Real run; the commit is dated 2026-09-30 through GIT_AUTHOR_DATE/GIT_COMMITTER_DATE, and the duct log names shown carry the same date. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
The graphic is cropped to end below the result tiles. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
… later Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
It was a rounding bound from ps printing pcpu to one decimal, not a measured spread (one run per phase), and read like an error bar. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
… easy Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
"You drive, the agent reads the records directly" replaces the heading, the bold lead line and the sentence under it, which said the same thing. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
Height cap 300 -> 360 px, into the space the shorter You drive panel freed. Co-Authored-By: Claude Code 2.1.299 / Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Keep the command as recorded. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Swap the left and center columns; each keeps its width. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The original column order reads better. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ted label Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
One-line speaker labels, the agent's con-duct ls / tail calls, and a lead-in line. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Nothing is marked simulated any more; the Claude Design graphic is now in the left column. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Invisible links (inherit colour, no underline) so print is unchanged and the PDF is clickable. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Full / half / empty circles (green / amber / grey) replace the check, cross and star; fixed layout gives the five tool columns equal widths. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Member
|
Is this the final version we should merge @asmacdo which you submitted for print??? |
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
It never printed (display: none under @media print); the 30 dpi render of a re-export is pixel-identical to the print file. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
It is the print source, not a mockup; the name now matches duct-poster.pdf. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The exact file sent to posternerd and published on Zenodo (10.5281/zenodo.23401552), md5 699f5a5711ec3559855a15bba340f3f1. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The source image was generated with ChatGPT by Austin Macdonald for the duct blog post and traced in Inkscape (con/duct#207), so it is not third-party material of unknown origin; it falls under the catch-all with the CON marks. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Member
Author
|
Yes, this is final:
🤖 Generated with Claude Code |
asmacdo
marked this pull request as ready for review
October 15, 2026 20:48
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The con-duct poster for the US-RSE'26 poster session (Oct 19–21, San Jose): one HTML page exported to a one-page 42×36 in PDF. The committed PDF is the exact file sent to the printer and published on Zenodo: https://doi.org/10.5281/zenodo.23401552 (version 2 of the abstract's record).
What's in it
2026-usrse/duct-poster.html: the poster source. Export withchromium-browser --headless --disable-gpu --no-pdf-header-footer --virtual-time-budget=5000 --print-to-pdf=duct-poster.pdf "file://$PWD/duct-poster.html"from2026-usrse/(needs network for IBM Plex from Google Fonts).2026-usrse/duct-poster.pdf(annexed): the print file, md5699f5a5711ec3559855a15bba340f3f1.2026-usrse/duct-poster-assets/: duct, CON and Dartmouth (D-Pine) logos, QR codes (con/duct and the abstract's DOI 10.5281/zenodo.22833931), the fMRIPrep plot from con/duct-gallery.REUSE.toml: D-Pine under the third-party logo marker; the duct logo stays under the catch-all with the CON marks;uvx reuse lintpasses.git-annex
Annexed content is on https://hub.centerforopenneuroscience.org/con/talks (an anonymous clone gets every file), not yet on datasets.datalad.org (I have no falkor access). annex-check fails anyway: its hub remote never resolves from GitHub's runners, on any PR. It goes green once the content is copied to con-site or the action reaches the hub.
Verified
Notes for reviewers
con-duct lsandtaillines are real commands, the conversation around them is written for the poster. Every other command output is from a real run.🤖 Generated with Claude Code