Eve library

Tara catalog and records · delivery receipt · 8 September 2026

Guide

Revision 1

On this page

The 22-slide guide and PDF follow one recommendation into linked practice, content and teacher records. Native containment and branch diagrams, an ownership exchange, record specimens and worked comparisons explain exact construction behavior before the independent validation boundary. Authoring records the sequence and source findings. Implementation baseline: 5532d1513abcccb25a181c9c7a4e187a8b936536; source pin: 1ceae52c7896a59ac74a4429d410f3218f0b91c6.

All 680 earlier slides, scripts, narration entries and MP3 hashes remain exact. Twenty-two new tracks add 1,742.993 seconds (29.0 minutes). The library has 66 guides, 702 unique slides, 725 guide-slide instances and 403 cited paths. The combined Tara feature units and full V1 program remain open.

Evidence#

  • Twenty source cases execute complete pinned domain models, type helpers, three actual feature taxonomy modules and independent contract modules with local Zod. The feature facade exposes those three modules; it does not execute unrelated barrel exports. Declared inputs characterize references, repeated inputs, normalization, metadata propagation, cadence and validation. Source continuity binds twelve cited or transitively loaded paths to both revisions.
  • The settled presentation and preservation suite passes 108 cases. Targeted ESLint, Ruff, formatting, generated freshness and coverage checks pass. Runtime product code is unchanged; no broad product build gate was needed.
  • The browser matrix passes 160 checks over desktop, compact desktop, 390/320-pixel layouts, three center embeddings and six enlarged-text cases. It retains 88 captures, 66 embedded accessibility checks and three deliberately rejected removals: breathing array, breathing branch and duration invariant. No fit finding or browser runtime error remains.
  • The focused center review passes fourteen checks, preserves all 65 earlier guide HTML files and four shared UI assets, checks navigation, search, history, mobile focus and both audio boundaries, retains 44 embedded captures and exports all 22 PDF pages. Browser audio output was disabled.
  • The PDF text report verifies exact visible-body labels and page order. Optical review records actual inspected images. All compact draft pages and selected 320-pixel, final embedded and PDF images were inspected; text extraction is not a visual-legibility claim.
  • New audio receives pause/encoded-audio validation and final scoped decode. Speech continuity compares all 680 earlier entries, scripts and MP3 hashes to their published evidence. All 702 current hashes are bound by this receipt.

Corrections and reproduction#

The initial containment authoring lacked explicit contains edge kinds and was corrected before the successful draft render. Optical review replaced an unclear space-marker glyph with quoted input and removed an implied upper count ceiling. Narrow review found one encoded route overflowing at 320 pixels; separating its ID from its prefix retains both pieces and passes. The settled comparison headings describe their content and peer rows have equal emphasis. The highlighted adapter is an explicit owner selection. Failed draft/diagnostic logs are retained in logs and attempts without counting them as passing delivery. An initial generic ESLint invocation lacked the owning presentation configuration; the correctly scoped invocation passes unchanged.

Run python3 docs/presentations/presentation-center/verification/v1-tara-catalog-2026-09-08/verify-receipt.py from the repository root. It binds authored/generated artifacts, evidence and all recording hashes. Later library changes invalidate this frozen receipt; use its immutable publication revision for historical verification.

The full V1 program remains active, including recommendation/search/saved/launch consumers, full taxonomy selection, native callers, companions, mentors and workbenches. Human comprehension, normal-speed listening, actual member/database/media/device and deployed acceptance were not performed. Source-constructor metadata does not establish those outcomes.