# v0.18.0 Research Accessibility, Print, and Long-Form QA Validation ## Scope This validation covers the v0.18.0 accessibility, print, and long-form reading quality pass across the public Research Navigator, Research Library and all 25 report readers, Research Methodology, References & Source Discovery, six research topic hubs, and the Glossary. The work refines presentation and interaction semantics without changing the durable research corpus, report manifest, accepted glossary definitions, topic membership, or v0.17.0 Navigator evidence boundary. No live deployment was performed. ## Accessibility semantics and interaction checks Static and HTTP-rendered checks completed with zero errors across the canonical route set. Verified behavior includes: - one visible document H1 per canonical route; - no duplicate rendered IDs in the checked route set; - all checked ARIA ID references resolve to rendered targets; - images on checked canonical routes have alt attributes; - search controls on the audited research surfaces have accessible labels; - buttons on the audited research surfaces have accessible names; - Research Library, References, Glossary, and Navigator result/status feedback uses explicit status/live-region semantics where dynamic result feedback is useful; - report truth-boundary, source/review, and citation/link controls have explicit group labels; - external reference links that open a new tab include an assistive-technology-only new-tab announcement; - fenced-code copy controls have explicit accessible names and atomic status feedback when code blocks are present; - the Navigator `/` shortcut does not capture the key while focus is in editable controls, links, buttons, disclosure summaries, or button-like interactive elements; - research filter controls and disclosure summaries meet the v0.18.0 first-party minimum target-size rules; - reduced-motion preferences suppress report reading-progress behavior and remove the audited research-surface transitions/transforms that are not necessary for meaning. Native JavaScript passed syntax checking. Direct browser/assistive-technology automation was not established in this environment, so keyboard execution, focus behavior, announcement timing, and screen-reader output are supported by static code and semantics review rather than claimed field-lab certification. ## Long-form and narrow-screen checks The report-reader contract from prior releases remains in force: constrained reading width, responsive table wrappers, native disclosure table of contents, stable section permalinks, source/review identity, related research, adjacent report navigation, and glossary bridges. The v0.18.0 pass adds defensive long-content behavior for audited research surfaces: - long URLs and unbroken content may wrap instead of forcing page-level horizontal overflow; - code/preformatted content is bounded to its reading surface; - report tables retain their responsive screen presentation and include accessible region semantics plus keyboard reachability; - 82 rendered report tables were checked, including 2,969 labeled body cells used by the narrow-screen card presentation; - 699 report section permalinks were rechecked; - 530 route-level HTML fragment links were resolved during the canonical-route sweep. ## Print contract First-party print CSS is now defined for the long-form report readers and Research Methodology page. Print presentation intentionally removes interactive and decorative site chrome while retaining the material needed to understand and verify the printed document. For report readers, the print contract preserves: - report title and substantive report content; - the visible research/truth boundary; - source attachment identity; - source and curated SHA-256 values; - durable source/review links where useful; - semantic tables rather than the narrow-screen card transformation; - wrapped preformatted/code content so it is not intentionally clipped by the print stylesheet; - useful hyperlink destinations for relevant report/methodology links. The stylesheet removes header/footer navigation, breadcrumbs, reading progress, interactive table-of-contents chrome, copy controls, related/discovery modules, and other nonessential interaction UI from printed output. Methodology substantive sections remain printable under the same document-oriented rules. No multi-browser printer or PDF-engine matrix was available, so this establishes the repository print contract and static CSS behavior rather than guaranteeing identical pagination in every browser/operating-system print engine. ## Navigator query-state and canonical-invariance regression The v0.17.0 shareable-state boundary was rechecked with clean and variant Navigator requests. Verified behavior includes: - optional `q` and `type` parameters do not alter the canonical Navigator URL; - Open Graph URL remains the query-free canonical Navigator URL; - JSON-LD remains query-independent and parseable; - title and meta description remain query-independent; - markup-like query input is escaped in visible output; - invalid result types fall back to the accepted all-results state; - visible query state remains bounded; - all server-rendered Navigator groups remain reachable without JavaScript. The v0.18.0 shortcut guard additionally avoids claiming `/` when the event target is an input, textarea, select, link, button, disclosure summary, button-role element, or editable content. ## Full repository checks The focused repository validation completed with zero errors: - 106 PHP files passed `php -l`; - native JavaScript passed `node --check`; - 85 canonical sitemap routes returned HTTP 200; - all 85 routes have unique titles and unique meta descriptions; - all checked canonical and Open Graph URLs match their clean route identity; - one parseable JSON-LD graph is present on every canonical route; - 182 missing-slash, explicit route `index.php`, homepage `index.php`, and legacy root-PHP redirect assertions passed; - 153 unique rendered local target paths were checked; - 530 route-level HTML fragment links resolved; - 25 report readers were checked; - 82 report tables and 2,969 labeled responsive table cells were checked; - 699 report section permalinks were checked; - 804 outbound reference links were checked for the first-party new-tab accessibility announcement; - 201 repository-local Markdown links resolved; - 67 durable UAIX pointer records and 91 declared section/review destinations resolved; - 212 conservative explicit UAIX/repository path references resolved; - the exact required UAIX setup/update URL remains present once in each required owning record; - handoff intake contains inert `.keep` placeholders only; - active taboo-rule leakage is zero; - lightweight private-key/token-marker scanning found zero candidates. ## Research-authority preservation All 25 manifest-backed curated report bodies are byte-identical to v0.17.0. All 25 curated SHA-256 values match `report-manifest.json`, and all 25 original source attachments available in the task environment match their recorded source SHA-256 identities. Every curated report retains the required `#curation-boundary` and `#research-body` anchors. `docs/long-term-memory/research/report-manifest.json` is byte-identical to v0.17.0, and the accepted glossary registry is byte-identical to v0.17.0. The accessibility/print pass changes presentation semantics only; it does not add research claims, legal conclusions, topic membership, glossary definitions, citation authority, or external-source verification. ## External and production state not established This validation is not an external accessibility certification and does not establish results from a real assistive-technology/browser lab, mobile-device lab, printer/PDF-engine matrix, production host, or field-performance environment. It also does not establish production DNS/TLS/rewrite/cache behavior, search-engine indexing, external-reference reachability/currentness/authority, current legal/scientific facts, or independently verifiable machine-execution provenance. Those remain separate deployment, external-verification, or field-testing concerns. ## Packaging note The release is packaged for direct web-root extraction. Final archive clean-extraction checks and archive SHA-256 are performed after this report is written; the hash is supplied in the release handoff rather than embedded here so recording it cannot mutate the archive it describes.