<a id="v0240-report-reference-context-validation"></a>
# v0.24.0 report-reference context validation

Date: 2026-08-12  
Release: 0.24.0

## Scope

This validation covers the v0.24.0 per-report external-reference context and section-traceability pass while preserving the accepted v0.23.0 compact global References representation, v0.22.0 delivery/cache contract, 109-route public surface, 49-report research authority, 42-entry source-bundle ledger, responsive media, accessibility/print behavior, clean URLs, and UAIX/taboo boundaries.

## Report-local reference traceability

The local extractor still recomputes exactly:

- **49** curated reports;
- **3,932** external URL occurrences;
- **1,900** unique normalized external URLs;
- **1,070** normalized domain groups;
- **2,014** unique normalized-URL-to-report relationships.

For every report, the detailed line-aware extractor was compared to the existing report-local URL extractor and required to produce the identical normalized URL multiset and occurrence count. The detailed layer adds only nearest-preceding rendered-heading context; it does not change URL normalization or global evidence totals.

Rendered report-reader checks require:

- exactly one **References in this report** disclosure on each of the **49** report readers;
- **2,014** report-local unique URL rows across the corpus, exactly matching the unique URL-to-report relationship count;
- exact report-local occurrence counts on every URL row;
- a deterministic section-key legend generated from the same shared collision-safe `md_heading_id()` helper used by report TOCs and section permalinks;
- every emitted section fragment to exist in that rendered report;
- every report-local domain backlink to resolve to one of the existing **1,070** stable fragments on `/research/references/`;
- explicit zero-reference presentation for the **4** curated reports with no extracted external HTTP/HTTPS URLs;
- server-rendered/no-JavaScript access to all report-local reference rows and section/global-index links.

Across the rendered report-local rows, **2,019** section links are emitted. Section association remains lexical/structural occurrence evidence only. It does not establish that an external source supports every statement in a section, or that the source is current, correct, reachable, authoritative, or endorsed.

## Global References regression

The compact `/research/references/` representation remains unchanged from v0.23.0:

- raw HTML: **772,144 bytes**;
- **3,932** occurrences;
- **1,900** unique normalized URLs;
- **1,070** stable domain fragments;
- **2,014** URL-to-report relationships.

The v0.23.0 67.1% raw-HTML compaction gain is therefore preserved.

## Report-reader local performance budget

Deterministic v0.23.0-to-v0.24.0 raw response comparison across all 49 report readers:

- v0.23.0 report readers: **62,112–162,463 bytes**, mean **112,438.8 bytes**;
- v0.24.0 report readers: **63,078–198,320 bytes**, mean **136,276.7 bytes**;
- added report-local traceability markup: **966–42,483 bytes**, mean **23,837.9 bytes**;
- largest percentage growth: **44.7%**;
- largest resulting report reader remains below **200 KB** raw HTML in the local renderer.

Human-readable budget: `../performance/v0.24.0-report-reference-context-budget.md#v0240-report-reference-context-budget`  
Machine-readable budget: `../performance/v0.24.0-report-reference-context-budget.json`

These byte measurements are repository-local comparison evidence, not field performance or deployed network-transfer measurements.

## Public route, metadata, and delivery regression

The canonical sitemap remains **109** clean public routes with **8** image entries and **1** video entry. The regression suite checks all canonical routes for HTTP 200, one H1/main/footer, unique non-empty titles and descriptions, self-canonical identity, matching Open Graph URL, parseable JSON-LD, unique IDs/resolving ARIA references, named controls/image alternatives, release-aware `/assets/css/site.css?v=0.24.0` and `/assets/js/site.js?v=0.24.0`, and absence of local public `.php` navigation targets.

Clean-URL coverage remains **230 redirect assertions**: 108 missing-trailing-slash variants, 108 clean-route `/index.php` variants, root `/index.php`, and 13 retained legacy root PHP endpoints.

The v0.22.0 optional Apache delivery configuration remains guarded with `<IfModule>` and production module/cache/compression behavior remains unverified.

## Research authority preservation

No curated research body or accepted glossary definition changed in this release.

- **49/49** curated report SHA-256 values match `report-manifest.json`;
- **49/49** original source SHA-256 values re-match available supplied task evidence by hash;
- all 49 reports remain byte-identical to the accepted v0.23.0 package;
- every curated report retains `#curation-boundary` and `#research-body`;
- `report-manifest.json`, `includes/glossary-data.php`, and `source-bundle-intake-2026-08-12.json` remain byte-identical to v0.23.0;
- the source-bundle ledger remains exactly **42 entries: 24 new-curated + 18 duplicate-existing**.

## Syntax, links, UAIX, and hygiene

The final validation cycle passes:

- **130/130 PHP files** through `php -l`;
- native JavaScript syntax through `node --check`;
- **7,738** rendered local path/fragment target combinations;
- **316** repository-local Markdown links;
- **101** durable UAIX pointer records with **100** declared section checks;
- **325** explicit repository-path references in active UAIX records;
- **46** rendered responsive-candidate occurrences representing **24** unique local candidate path/descriptor pairs, with declared widths checked against actual image widths;
- the exact UAIX update URL once in each required owning file;
- only inert `.keep` files in the two active handoff buckets;
- zero taboo leakage outside the owning taboo record;
- zero findings in the focused raw-secret marker scan.

## Evidence limitations

This is repository/local-server validation. It does **not** establish external reference reachability, source correctness/authority/currentness, production Apache modules, HTTP compression, cache behavior, browser/network timing, Core Web Vitals, live DNS/TLS/redirect state, indexing, webmaster-console state, accessibility certification, or independently verifiable machine-execution provenance.

No live deployment was performed.

## Next bounded action

Create a deterministic release-integrity manifest plus a public Transparency integrity page that lets readers recompute package-file hashes while explicitly separating package-byte integrity from live deployment, authorship, external factual correctness, or autonomous execution. The full bounded task is stored in `.uai/next-recursive-prompt.uai`.

## Root-package validation

A candidate root archive was clean-extracted into a separate directory and the same full validation suite passed again against the extracted bytes: 130 PHP files, 109 canonical routes, 230 redirect assertions, all report-local reference/section checks, global References totals/size, all 49 research authority hashes, bundle-ledger preservation, local links/fragments, UAIX pointers, taboo containment, secret hygiene, and media/delivery regressions. The archive uses direct web-root shape with no enclosing project directory. The final ZIP is rebuilt only after this validation report write-back and is re-extracted/rechecked before handoff; its final SHA-256 is supplied with the release handoff rather than self-recorded inside the archive.
