One-column reflow, 48px route headlines where specified, mobile navigation, large disclosure targets and no page panning.
Access proofbook · tested behavior and limits
The argument should remain usable without motion, a mouse or a visual chart.
This preview uses semantic HTML, visible focus, native controls, text alternatives and reduced-motion styles. Browser and assistive-technology testing remains an ongoing release requirement.
Conformance boundary
Reproduced behavior is not certification.
This ledger records app-local implementation and reproduced browser behavior. It is not a WCAG conformance claim, certification, independent audit or guarantee across every browser, operating system, device or assistive technology.
Human audit gated10 behavior records · boundaries attached
Inspect what was tested and what it cannot prove.
Open one record at a time. The URL stores reading focus only; it does not change conformance, hide the static record or create a report.
Input02reproduced localCore reading controls operate from the keyboard
App-local claim
Navigation disclosures, native details, priority sliders and editorial lenses support keyboard activation; relevant lens groups also support directional-key focus movement.
- Evidence
- Local browser QA has reproduced Enter, Space, Escape and arrow-key journeys on the mobile menu, revision details, priority controls and dedicated editorial routes.
- Reproduction method
- Complete each primary journey without pointer input, observe URL and live-status changes, and verify the focused control remains identifiable after the action.
- Boundary
- Reproduced journeys do not cover every browser shortcut, switch-control configuration, voice-control grammar or alternative input device.
Focus03reproduced localFocus is visible and returns from temporary surfaces
App-local claim
The global focus-visible treatment uses a high-contrast outline. The mobile menu returns focus on Escape, and the source dialog moves focus to Close then returns it to the invoking citation.
- Evidence
- CSS inspection confirms a three-pixel focus outline with offset. Browser QA reproduced menu focus retention and source-drawer focus entry and return.
- Reproduction method
- Tab through representative pages, open and close each temporary surface with the keyboard, and inspect activeElement before and after dismissal.
- Boundary
- Visible focus styling does not prove that focus order is ideal in every long article or that all third-party browser modes render the same outline.
Status04reproduced localDynamic reading changes announce status and failure
App-local claim
Weighted comparison, research-priority and editorial lens controls use polite live outputs for state changes, malformed-link recovery and explicit no-result boundaries.
- Evidence
- Regression tests require an aria-live summary, under-these-weights language and all-zero suppression. Browser QA reproduced status after weight, priority, source and malformed-state journeys.
- Reproduction method
- Trigger valid, all-zero and malformed states, inspect the live output and confirm it identifies what changed while preserving the legal-quality boundary.
- Boundary
- DOM live-region updates have not yet been independently reviewed across NVDA, JAWS, VoiceOver and TalkBack speech settings.
Evidence05reproduced localOfficial-source detail remains reachable without losing place
App-local claim
Source citations are ordinary links with an enhanced native dialog when supported; the fallback route opens the source ledger when showModal is unavailable.
- Evidence
- SourceDrawer uses a labelled dialog, Close control, Escape behavior and stored trigger focus. Browser QA has opened Indian and international source records and verified focus return.
- Reproduction method
- Open a citation with the keyboard, inspect the dialog name and status, close with Escape, verify focus return, then inspect the non-dialog ledger link.
- Boundary
- The publication cannot guarantee the accessibility, language, tagging or reading order of linked external government pages and PDFs.
Reflow06reproduced localThe page shell reflows without page-level horizontal panning
App-local claim
Representative production routes fit 320, 768 and 1440 pixel viewports; wide legal tables retain intentional local scrolling rather than widening the entire page.
- Evidence
- Repeated browser QA compares body scroll width with the viewport after every responsive change and records page-specific grid and target dimensions.
- Reproduction method
- Inspect representative routes at the three required widths, attempt horizontal page movement and separately verify any intentionally scrollable table region.
- Boundary
- Fixed viewport checks are not a substitute for a complete 200 and 400 percent zoom, text-spacing and browser-font-override matrix.
Motion07implemented reproducedReduced-motion preference removes non-essential movement
App-local claim
The shared reduced-motion media query disables smooth scrolling and reduces animation and transition duration globally without hiding content.
- Evidence
- The production CSS includes a prefers-reduced-motion rule, and prior browser emulation reproduced auto scroll behavior and the 0.01 millisecond fallback duration.
- Reproduction method
- Emulate reduce, inspect the matched media query and computed scroll, animation and transition behavior on representative interactive routes.
- Boundary
- The rule is an implementation safeguard, not evidence about vestibular comfort for every user or every future animation.
Resilience08implemented reproducedCore editorial evidence survives without scripting
App-local claim
Progressive lens controls begin hidden while substantive records remain in static HTML; noscript notes explain when optional focus controls are unavailable.
- Evidence
- Integrity tests inspect hidden-control and visible-record patterns. The static Astro build contains the full article, source and revision content before client scripts run.
- Reproduction method
- Inspect generated HTML with scripts unavailable and confirm evidence records, source links, empty states and ordinary navigation remain present and understandable.
- Boundary
- Copy-link, live filtering, dialog enhancement and lens focus require scripting; their underlying content and destination links remain available.
Wayfinding09reproduced localThe publication index supports search and directional-key movement
App-local claim
The global research index uses a labelled native dialog, a progressively enhanced methodology link, a local search field, live result counts and Arrow, Home, End and Escape keyboard behavior.
- Evidence
- The static component contains every indexed destination as an ordinary internal link. Browser QA opens it by trigger and slash shortcut, filters results, moves focus through visible matches and returns focus after close.
- Reproduction method
- Open the index from the header and with slash, type a known and unknown query, use ArrowDown, ArrowUp, Home and End across visible links, close with Escape and verify focus returns to the trigger.
- Boundary
- A searchable route map improves wayfinding but does not prove that every label, synonym, focus order or result description works equally well for every reader and assistive technology.
Data view10reproduced localWeighted outcomes have semantic, announced and printable alternatives
App-local claim
The decorative constellation is paired with a live text summary, score and missing-data table, supporting/counterargument records and a sensitivity table; no result depends on colour or motion alone.
- Evidence
- Playwright verifies visible maths, URL restoration, all-zero behavior, source inspection and sensitivity updates. Axe reports no serious or critical findings on the comparison route.
- Reproduction method
- Change a weight by keyboard, read the live summary and table, inspect a source, emulate reduced motion and print, then reopen the copied URL.
- Boundary
- Automated and local browser evidence does not replace an independent screen-reader and print-production review.
Five known limitations
Unknown stays visible.
No limitation is converted into a passed check by design polish, automated output or the absence of a reported issue.
No independent assistive-technology audit
No qualified external reviewer or verified user study has tested the complete publication with NVDA, JAWS, VoiceOver, TalkBack, switch control or voice control.
- Impact
- The current evidence supports implementation confidence only; it cannot support certification or a universal usability claim.
- Evidence needed
- A documented human audit across agreed browsers, operating systems, screen readers and input modes.
Zoom and text-spacing matrix remains incomplete
Required viewport checks exercise narrow reflow, but a route-by-route matrix at 200 and 400 percent zoom, increased text spacing and browser font overrides has not been completed.
- Impact
- Long labels, data tables or dense evidence records may expose issues not visible in the fixed-width checks.
- Evidence needed
- Repeatable zoom, reflow and text-spacing checks on every primary journey and long-form route.
External official sources are outside app control
Government pages and PDFs linked by the source ledger may have inaccessible document tagging, reading order, language or interaction patterns.
- Impact
- The source drawer improves context but cannot repair the linked authority's document or website.
- Evidence needed
- Record source-specific barriers and provide a legally faithful accessible explanation without replacing the controlling record.
No public accessibility feedback route
The preview has no form, inbox integration, account, messaging endpoint or server-side accessibility intake.
- Impact
- Readers cannot yet report a barrier through the site, and no response time or case-retention promise is made.
- Evidence needed
- An approved human owner, minimal collection fields, adjacent notice, response standard and retention schedule.
Contrast and forced-color review is not independently signed off
Visible focus and a forced-colors rule are implemented, but no independent contrast audit or high-contrast user review has been supplied for every component and state.
- Impact
- Design intent and automated inspection cannot substitute for state-by-state human review across display modes.
- Evidence needed
- Component-level contrast measurement plus human review in Windows High Contrast and other forced-color modes.
Release controls
One local baseline. Five gates remain.
Approve a human owner, purpose-specific contact route, adjacent collection notice, response standard and retention schedule before collecting accessibility feedback.
Local baseline
Keyboard, responsive and static behavior is reproducible.
App-local integrity, build and browser checks cover the stated routes and viewport matrix.
Reproduced locallyHuman audit
Assistive-technology review remains required.
No independent cross-browser screen-reader and alternative-input report has been supplied.
GatedReflow audit
Zoom and text-spacing coverage remains required.
Fixed viewport QA is complete; the full enlargement and override matrix is not.
GatedVisual audit
Contrast and forced-color sign-off remains required.
Implementation rules exist without independent state-by-state approval.
GatedSource access
External-source barriers need a publication policy.
The site links controlling official records whose accessibility is outside app control.
GatedFeedback ownership
A human-owned accessible contact route remains required.
No approved owner, personal-data fields, retention schedule or response standard exists.
GatedFeedback intake
No accessibility submission route is active.
There is no accessibility feedback form, inbox integration, account, message route or server-side intake in this private preview.
Inspect correction controlsTrust the boundary
Implementation evidence can be checked. Human experience still has to be heard.
This proofbook stays deliberately incomplete until independent review and a human-owned feedback route exist.
Article revision record
What changed on this page.
This register begins with the private-preview editorial rebuild. It records material content, method, interface and trust-policy changes without inventing a history that predates the available repository evidence.
ROUTE/accessibility
ENTRIES03
Inspect the full revision register Presentation or interaction changed without changing a legal propositionAdded a ninth behavior record for the global research index, covering its native dialog, progressive fallback, live result count, directional-key movement and focus return.
- Why this changed
- The new publication-wide wayfinding surface needs an explicit reproduction method and limitation instead of being implied by generic keyboard and focus claims.
- Claim impact
- No WCAG conformance or certification claim was added; the record documents app-local behavior while preserving independent assistive-technology review as a release gate.
- Review state
- Human legal or editorial review still required
- Change ID
ACCESS-20260828-03
Source impact
No legal source changed; this entry records method, interface or trust policy.
Publication, privacy, correction, accessibility or use policy changedReplaced the generic accessibility statement with a versioned proofbook covering eight reproduced behaviors, five known limitations and six release gates.
- Why this changed
- A credible accessibility statement must separate app-local evidence from unverified assistive-technology, zoom, contrast, external-source and feedback-channel coverage.
- Claim impact
- No WCAG certification, independent audit or universal usability claim was added; a public feedback route and five human review gates remain disabled or outstanding.
- Review state
- Human legal or editorial review still required
- Change ID
ACCESS-20260826-02
Source impact
No legal source changed; this entry records method, interface or trust policy.
Publication, privacy, correction, accessibility or use policy changedDocumented keyboard operation, focus treatment, reduced-motion behavior and the private preview's current accessibility boundary.
- Why this changed
- Accessibility claims should describe reproduced behavior and known limits rather than imply certification or universal conformance.
- Claim impact
- The statement remains an implementation record, not an accessibility certification or external audit result.
- Review state
- Human legal or editorial review still required
- Change ID
ACCESS-20260821-01
Source impact
No legal source changed; this entry records method, interface or trust policy.