Eve library

Tara web search and result navigation · Tara · find a practice, chapter 3 · verification receipt · 11 September 2026

Guide

Revision 1

On this page

Scope: the third chapter of the Tara · find a practice track. Sixteen inherited slides became twenty-two: a cover, four section dividers, sixteen content slides and a close, in four sections of four. The brief records what the chapter is for and why each treatment was chosen; the track brief is tara-find-a-practice-track.md; state.json holds the exact counts, tool results and findings.

What was checked#

  • Build freshness. build-eve-oshun.py --check passes on the delivered tree. The chapter is 22 slides running 28.4 minutes, with no table in it.
  • Two captures, both already pinned. The inherited chapter had two product-view slides citing images from the search harness. They are now capture-callouts with numbered pins and a legend, over the same bytes at the same revision. The first is the summary line for calm — the settled query, the clear control and the seven-result count with its synonym flag. The second is one complete result row that is in the list without containing the word that was typed.
  • Every treatment was re-decided. The plan's card-grid default was applied where the rows really are peers and replaced where they are not: a state machine for the three settled states, a stat panel for the inventory counts, two compare panels with verdicts, a graph diagram for the events that write history, and two decision trees that kept two outcomes each because the source has no third branch.
  • Bindings. No assignment file and no teaching guard binds a slide in this chapter; every assignments/*.json evidence entry was checked by id and the node suite is green on the rewritten manuscript.
  • Browser rendering. eve-edition-review.mjs --guide tara-web-search sweeps all 22 slides at 1440×900, 1280×720 with audio chrome, 390×844 and in print — 88 measurements, no spill, no document overflow, no page errors, no findings.
  • Narration. Twenty-two tracks rendered on this host's CPU in batches of six and verified; the library is complete at 1924.4 minutes with nothing pending.
  • PDF. 22 pages, exact authored visual text on every one, in order, read back from the downloadable decks/tara-web-search.pdf.
  • Suites and lint. Python 473 passed, 0 failed. Node 489 of 489. ruff check and Prettier clean.

What the chapter names#

None of this is new to the code; what the rewrite adds is saying it as one argument rather than sixteen facts:

  • Expansion runs one pass over the words the member typed, so stress adds calm and stops there, while typing calm reaches both groups.
  • Popularity boosts are gated behind a non-zero text match, so a highly rated record with no textual match scores zero rather than a small number.
  • The sorted ranking is discarded at render time in favour of a fixed type order, so the page cannot be read downwards as a relevance list.
  • Highlighting receives the original words rather than the expanded ones, so a row can be present for a reason the page never shows — which is the second capture.
  • History is written before the route is consulted, and the route returns nothing for meditations, which are the largest kind in the inventory and the first section on the page. The ordinary click on this screen records a query and goes nowhere.

What was fixed#

  • The inventory slide broke twice. As the planned containment graph it was four rows tall and ran 49.6px past the page at 1440×900. As a stat panel it then pushed the page 124px sideways at 390px, because a stat panel's label names the file every number was read from and a repository path has no break opportunity in it. .specimen-label now wraps anywhere, which is a fix for every stat panel rather than for this one.
  • test_eve_design held the containment parentage contract on a named slide, and that name had to move twice in two days as the chapters holding it were rewritten. It now selects whichever delivered slide still uses the layout and fails loudly if the library ever has none, so it stops moving.

What was not checked#

  • No human reader session and no listening review at normal speed.
  • Nothing about any destination. The chapter records that three result kinds can produce a route request; whether those pages render, resolve content or succeed is chapter four's subject and the destinations' own evidence.
  • The interaction and reduced-motion observations are exactly what they say and no more. They are not a screen-reader session, a target-size review or a contrast audit, and the slide that carries them says so.