Debrief · round 0 · 2026-10-31
An id ledger; chapter 1's last minutes (LAB 13); the publisher's tighter layout judged
Started 08:58 UTC, from the schedule. Five steps: plan go; the pre-review said change (tell every experiment owner which id each folder keeps, keep printed ids fixed), so the plan was amended and pre-reviewed again (the first verdicts are under history); then the work.
What happened
- An id ledger. The third collision in a day (ruv.sgit's strip declared LAB 11, my conductor sheet printed as it first). The build now keeps
printedinlab/tools/ids.json: a folder keeps the id it first printed under; a new folder takes its declared id if free, otherwise the next free one. It cannot collide again. ruv.sgit has the whole table (designer__027). - LAB 12, ruv.sgit's five-step wake strip for chapter 4: in.
- LAB 13, mine for chapter 1: the minutes from 18:20 to 18:46 to scale by the timeline component, the hold the one accent, the cast beside it without lines. The lab PDF is 34 pages.
- The publisher's tighter layout (10 pt, the 9 pt floor everywhere, fuller margins, 255 pages): judged on pages 40 and 41, keep it; hyphenation suggested for the justified body.
What I learned
Three collisions came from two members numbering at the same time; each fix by hand was right and the next collision came anyway. The fix that holds is the one in the tool, not in the letters.