Skip to content

Cut the Time chapter's repeated material - #31

Merged
mikemaccana merged 1 commit into
mainfrom
claude/tighten-time-chapter
Aug 7, 2026
Merged

mikemaccana merged 1 commit into
mainfrom
claude/tighten-time-chapter

Conversation

@mikemaccana

Copy link
Copy Markdown
Collaborator

Flat off main, independent of #30. 2,257 words to 2,023.

The chapter had four structures all summarizing the same material: an opening table, a freshness table, a closing table, and What You Learnt. Three of the four overlapped.

What went

  • The opening table. Its two rows restated, nearly word for word, the two sections directly beneath them — "counts block opportunities, roughly 400ms apart" and "only goes up, at a rate nobody can meaningfully bend" both appeared twice within ten lines. The two-clocks figure moves up into its place, so the opening now shows the difference rather than tabulating what the next page says.

  • Two columns of the closing table. "If the seconds are nudged" and "if the chain halts" recapped the halt section wholesale, and three of six rows said only "nothing to bend". Two facts in those columns appeared nowhere else — leverage multiplying a bad price, and the order book resuming untouched — and both move into the halt section, where the rest of that argument already lives. What remains is a three-column reference to what each program chose, which is the part a reader flips back to.

  • The duplicate two-clocks figure, which sat immediately under the decision tree offering a second answer to the question the tree had just answered.

  • The second reference to the Offchain Truth chapter, five lines after the first, and the second explanation of why the vault counts seconds, which the bullet in the timestamps section had already given.

  • "The factor advances by slots, not seconds", in a paragraph that opens by defining elapsed slots.

The one thing I kept that you flagged

You suggested killing the opening outright because "Measure money with the slot counter" duplicates Which Clock to Use for Each Job. I kept that sentence and cut the table under it instead. The sentence is the chapter's thesis in one line, and the tree is the full procedure for the cases the thesis does not cover — order of arrival, published units, human promises. Without it the chapter opens cold on a section heading. Happy to cut it if you'd rather the tree be the only answer.

Verification

npm run lint clean, node scripts/check-figure-geometry.mjs clean across 72 figures, npm run build-pdf at 243 pages.


Generated by Claude Code

The chapter said several things three times. Removed:

- The opening table, whose two rows restated, nearly word for word, the
  two sections immediately below it. The two-clocks figure moves up into
  its place, so the opening shows the difference instead of tabulating it.
- The second half of the closing table. Its "if the seconds are nudged"
  and "if the chain halts" columns recapped the halt section wholesale,
  and three of six rows said only "nothing to bend". The two facts those
  columns carried that appeared nowhere else, leverage multiplying a bad
  price and the order book resuming untouched, move into the halt section
  where the rest of that argument lives. What remains is a reference to
  what each program chose.
- The duplicate two-clocks figure, which sat under the decision tree
  giving a second answer to the question the tree had just answered.
- The second reference to the Offchain Truth chapter, and the second
  explanation of why the vault counts seconds.

2,257 words to 2,023.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ABFpszmPxRorwWtWGR2jXj
@mikemaccana
mikemaccana merged commit b1fac80 into main Aug 7, 2026
@mikemaccana
mikemaccana deleted the claude/tighten-time-chapter branch October 13, 2026 00:23
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants