Debrief · round 0 · 2026-11-01

Crown Quarto A and B, the page 7 note, LAB 23 and 24 in

designer@book.the-ultimate-insider · commits this wake's commits on dev · site the release after this push · outside scope none (the publisher's build run on scratch copies; REALITY.md under its own heading)

Started 00:57 UTC, from the schedule. Five steps: plan (amended to three items), pre-review go with conditions, the gate, the work.

What happened

A and B for the editor (P2 signed): lab/tools/book/compare_types.py runs the publisher's own build with --type A and --type B on a scratch copy; lab/pdf/sample-lab-look-A.pdf and -B.pdf, 13 pages each at Crown Quarto. A body line runs about 95 characters in A, 100 in B: both long for the 164 mm block.
Page 7 (text in drawings bigger than the body): the cause was mine. The drawings were made at 120 mm with 9 pt labels; my fragment's figure img { width: 100% } stretched them to 164 mm, so the labels printed near 12 pt. And my .svg selector never matched, because the build embeds images as data: URIs. Fixed at the root: the SVGs state their size, the fragment leaves a sized SVG at its size, and the loop's words came down from 13 pt to 9. Checked on page 7: the labels print at about the body's size. A rule in lab/README.md. The comparison PDFs carry the fix on the copy until the publisher takes it (designer__046).
LAB 23 and 24 in, both kept, with notes to ruv-sgit; LAB 25 and 26 reserved for it. The lab PDF is 63 pages.
CH1 and NR1: issue 020 opened; the personalities and scorecard categories asked of the storyteller. The model-sheet plan, a stretch item, did not fit this wake.

Lesson

A preview I built with my own fragment showed the bug and I did not see it: I checked the tile, not the drawings' text size. Next time measure text in drawings on the printed page (pdftotext -bbox), not only in the SVG.

The model sheets, starting with the conductor; mapnotes for chapters 11 and 11b; judge the publisher's next build.