clide/governance/decisions/accessibility.md

Accessibility + i18n Decisions

A11y + i18n are Tier-0 contracts, not Tier-6 polish.

---

D-20: A11y is a Tier-0 contract

• Date: 2026-04-21
• Decision: Every widget primitive wraps its interaction surface in a Semantics node at the point of creation. A11y coverage is a test-time gate (ci/test_a11y.sh), not a post-hoc polish pass. ensureSemantics() fires at app boot; Flutter's semantics tree is always populated.
• Rationale: Retrofitting a11y onto a grown UI is what every project that skips this promises to do later and then doesn't. Making it a Tier-0 contract costs one Semantics line per primitive and a semantic-coverage test; postponing costs a rewrite.
• Cost: Widget authors maintain correct labels; tests reject new primitives without semantics. Enforced by test/a11y/ coverage tests.
• Raised by: 2026-04-21 planning.

D-21: i18n is a Tier-0 contract (fframe pattern + locale-fallback chain)

• Date: 2026-04-21
• Decision: All user-facing strings resolve through a namespaced i18n catalogue loader ported from fframe's text-driven pattern, extended with a locale-fallback chain fframe lacks. JSON per locale; I18n.of(context).t('namespace.key', {vars}). Missing keys resolve down the chain (e.g. en_GB → en → default), never fail silently; missing at the base locale logs a dev-mode error.
• Rationale: Flutter's intl + ARB codegen is inflexible for plugin-contributed catalogs (see R-4) — we need per-extension catalogs that merge without a codegen step. fframe's shape fits; its silent-fallback behaviour does not, so we add the chain.
• Cost: JSON has no comments and no trailing commas; translation tooling has to accept that. Separate i18n facade on every feature.
• Raised by: 2026-04-21 planning.

D-22: WCAG-AA contrast gate on bundled themes

• Date: 2026-04-21
• Decision: Every bundled theme must pass a WCAG-AA contrast check on its canonical token pairs (text/background, link/background, focus-ring/background) at test time. ci/test_a11y.sh runs the gate; CI fails on regressions.
• Rationale: Themes drift under "looks nicer" tweaks; contrast regressions land silently. Running the gate on every PR is the cheapest insurance. Ran the gate on initial themes — caught one summer-night muted token at 2.81:1 (below AA), fixed before landing.
• Cost: Third-party themes (Tier 6) won't be gated until an extension-time test hook lands. Bundled themes are gated today.
• Raised by: 2026-04-21 planning. Refined by D-69 — the gate's strict pair set only applies to high-contrast variants; named themes keep their published palettes.

D-69: published themes are user contracts; ship -hc variants for a11y

• Date: 2026-05-17
• Decision: The four bundled themes that ship under a recognisable name — clide, midnight, paper, terminal — are user contracts. Their palette colours (including syntax tokens, status colours, and borderHi) MUST NOT be retuned to satisfy contrast gates. When a stricter contrast check would fail one of them, the fix is one of: (a) ship a sibling theme with -hc (high-contrast) or -cb (colour-blind) in the name and enforce the strict pair set only there, or (b) split canonicalPairs into a baseline set every theme must pass and an extended set that only the -hc/-cb variants must pass.
• Rationale: Users pick midnight because it looks like VS Code, paper because it reads as a drafting sheet, terminal because of the amber-on-near-black tmux feel. Quietly darkening paper's success/warning/info or boosting midnight's borderHi to pass a WCAG-AA check changes what they got and what they signed up for. A11y is a Tier-0 contract (D-20), but it's served by offering an accessible variant, not by overwriting the aesthetic ones. VS Code itself ships Default Dark+ and a separate Default High Contrast for exactly this reason.
• Cost: Two extra theme files per "named" theme when we add a11y variants. The bundled-theme contrast gate (D-22) needs a baseline/extended split so the named themes don't fail the strict pairs.
• Raised by: 2026-05-17 — user intervened mid-T-114 when I had retuned clide/midnight/paper/terminal palette entries to satisfy the expanded canonicalPairs; reverted, decision written, T-114 will follow this rule.

D-102: i18n routing — ext-id namespaces, core catalog, ClideSettings.i18n facade, contribution keys

• Date: 2026-06-19
• Decision: Implements D-21 across the whole app (epic T-462).
  • Namespaces: an extension's catalog namespace IS its id (builtin.<name>); the ExtensionManager eager-loads it on activation, so a built-in localizes with no hand-maintained registry. Framework chrome outside any extension (lib/widgets, lib/kernel, the shared reader chrome) resolves under one core namespace, preloaded at boot.
  • Read path: widgets resolve through the single D-101 facade — ClideSettings.i18n.string(context, key, namespace:, placeholder:) (+ .interpolated) — null-safe (returns the placeholder when no kernel is in scope, so primitives render in isolated tests).
  • Manifest labels: CommandContribution carries titleKey/i18nNamespace; the palette and menu resolve via a shared localizedCommandTitle, and the palette's fuzzy search matches the localized title. The settings schema carries i18nNamespace+titleKey on the category and labelKey/helpKey/option labelKey beneath it, threaded down by the renderer.
  • Storage: catalogs are bundled assets at assets/i18n/<locale>/<namespace>.json — the locale is a directory (en_us, future nl_nl, nl_be, en_eu, …), so a new language is a new folder of the same namespace files, no renames.
• Rationale: makes a complete translation set (e.g. a Dutch pack) a pure data drop — no code. ext-id namespaces need no registry; the facade keeps one widget-facing read path for theme/fonts/i18n (D-101); the locale-dir layout is cleaner to maintain and mirrors how an external extension ships its own catalog.
• Cost: every extension's manifest gains optional key fields, and framework primitives now depend on the (null-safe) facade. The pure-data search matcher (settingsFieldMatches) still matches the English label — display localizes, search-by-translation does not (acceptable refinement).
• Raised by: 2026-06-19, epic T-462 (i18n everywhere). Builds on D-101.

---