Remove terminal renderer references

CHANGELOG.md

@@ -12,6 +12,16 @@ independent of these package versions.
 
 ## Unreleased
 
+## v0.2.0-vm.15 - 2026-10-05
+
+The renderer examples now expose the supported web and desktop surfaces without
+shipping a terminal renderer.
+
+- Remove the retired terminal renderer examples, adapters, package entries and
+  dependency references.
+- Keep the browser and desktop semantic view contracts and their existing
+  qualification evidence.
+
 ## v0.2.0-vm.14 - 2026-09-30
 
 Native release checks keep the source checkout clean while testing archives

docs/malleable-roadmap-progress.md

@@ -81,12 +81,12 @@ this document records orchestration state.
 
 | Stage | Evidence | Delivery scope |
 | --- | --- | --- |
-| A | Qualified portable filesystem/memory lifecycle, 19 native/reference receipts and 38 cross-verifications; same task capture/actions across browser, desktop and TUI; shared session fixtures; actual renderer costs and limitations measured | No missing local exit evidence. Desktop's provisional CUA-inclusive action budget failed; performance remains experimental, with paint latency unisolated |
+| A | Qualified portable filesystem/memory lifecycle, 19 native/reference receipts and 38 cross-verifications; same task capture/actions across browser and desktop; shared session fixtures; actual renderer costs and limitations measured | No missing local exit evidence. Desktop's provisional CUA-inclusive action budget failed; performance remains experimental, with paint latency unisolated |
 | B | Qualified static fallback, independent mounts, keyboard continuity, expected-head rejection, preview/adoption and restoration; actual staging preview and owner browser checks pass | Staging qualification; production acceptance checks the deployed capture against the validated build |
 | C | Qualified ordered/deduplicated signals, same-capture explanations, persistent controls, journal/history and interrupted-publication/restart behavior | Local owner and public read-only captures share the same bounded read model |
 | D | Real Gateway attempts retained; independent pass/fail/inconclusive shadows; guarded adoption; one joined provider-reported USD 0.001369 debit, with the original unknown preserved | Missing/malformed provider billing remains unknown; recorded public fixtures make no inference calls |
 | E | Shared lifecycle in actual workerd, immutable core dependency pin, typed client, owner dashboard, A/A instrumentation, inert recovery and forward restoration | The [cloud delivery record](https://github.com/hraness/algal-cloud/blob/codex/application-roadmap/docs/application-roadmap.md) owns exact integration, isolated hosted qualification and dashboard evidence |
-| F | Final exact-source Dioxus/Ratatui package: 37 extraction checks, 11 packaged offline checks, actual native typing/select/draft interactions; real network-denied Apple revision change and matching packaged provider binaries | Functional exit qualified for macOS arm64; ad-hoc signed, not notarized; broader platforms and desktop performance are not production-qualified |
+| F | Final exact-source Dioxus package: 37 extraction checks, 11 packaged offline checks, actual native typing/select/draft interactions; real network-denied Apple revision change and matching packaged provider binaries | Functional exit qualified for macOS arm64; ad-hoc signed, not notarized; broader platforms and desktop performance are not production-qualified |
 | G | Qualified materially different task application, actual v1→v2 migration, workflow changes, typed forms/actions, core lifecycle conformance, pure evidence forks and explicit field conflicts | Full browser-local kernel explicitly deferred, not claimed by the browser renderer |
 
 Core commit `39607174063030eb4793ab65601a41d7d7c918d4` passed the aggregate
@@ -154,8 +154,8 @@ schema migration, explicit fork conflict resolution and separate durable session
 Its focused host suite passed 8 tests/81 assertions after independent review
 removed missing-evidence healing and cache masking. Nineteen complete
 native/reference receipts matched with 38 cross-runtime offline verifications.
-The Dioxus desktop and Ratatui builds passed. Browser, desktop and terminal
-share semantic capture/command data; five session-identity fixtures cover adding,
+The Dioxus desktop build passed. Browser and desktop share semantic
+capture/command data; five session-identity fixtures cover adding,
 changing, moving and removing controls while retaining drafts. Actual native
 interaction preserved a task and independent draft across schema upgrade and
 explicit rebase. It exposed Dioxus fast-typing loss and initial dropdown errors;
@@ -164,8 +164,8 @@ regressions exercise real virtual-DOM mutations, and actual final-package typing
 first-mount selection, panel remount, save and draft-reset checks passed.
 
 The final macOS arm64 archive passed 37 fresh-extraction checks and 11 checks
-under process-tree network denial. It includes a standalone expression-WASM host,
-Dioxus app and Ratatui executable; no checkout or Bun installation is required.
+under process-tree network denial. It includes a standalone expression-WASM host
+and Dioxus app; no checkout or Bun installation is required.
 The archive SHA-256 is
 `e59b81d33b20af0917da6211510e76d438155af6bc05abaa5f12c3e08707e556`;
 its clean source inventory SHA-256 is

docs/malleable-software-plan.md

@@ -131,16 +131,15 @@ Dioxus is a leading candidate, not a settled architectural dependency. Reviewed
 | Dioxus Web | WASM driving browser DOM; selectable mount root | Test as an embedded surface against a small semantic HTML baseline |
 | Dioxus desktop/mobile | Native Rust logic with a system WebView | Early local application candidate; do not call these platform-native widgets |
 | Dioxus Native / Blitz | Separate HTML/CSS/WGPU path; Blitz is beta | Later qualification, especially text input and accessibility |
-| Dioxus TUI | Deprecated | Use a separate Ratatui adapter |
 | Dioxus LiveView | Deprioritized server/WebSocket approach | Do not make it the foundation of offline operation |
 
-[Dioxus platform structure](https://dioxuslabs.com/learn/0.7/beyond/project_structure/), [Blitz status](https://github.com/DioxusLabs/blitz#status), [Ratatui's Elm architecture](https://ratatui.rs/concepts/application-patterns/the-elm-architecture/)
+[Dioxus platform structure](https://dioxuslabs.com/learn/0.7/beyond/project_structure/), [Blitz status](https://github.com/DioxusLabs/blitz#status)
 
 A compiled recursive renderer can interpret a changing ALGAL view tree using keyed components. Layout, content, tree structure and action bindings can change without compiling new Rust. Dioxus's RSX and VNode APIs support dynamic content, while optimized templates use static structures. Begin with a fixed interpreter over versioned widgets, and measure allocation/reconciliation behavior through repeated revisions. [RSX](https://dioxuslabs.com/learn/0.7/tutorial/rsx/), [template API](https://docs.rs/dioxus-core/0.7.10/dioxus_core/struct.Template.html)
 
 Development hot reload and experimental Rust hotpatching are useful tooling, but they are not ALGAL's production evolution protocol. Arbitrary new native code remains a separately built and admitted host extension. [Dioxus hot reload](https://dioxuslabs.com/learn/0.7/essentials/ui/hotreload/)
 
-The renderer spike should test one small interactive artifact in Dioxus Web, a desktop WebView and a minimal Ratatui renderer. Cover stable focus and draft preservation, IME input, keyboard navigation, accessible names, stale events, restart, repeated revision swaps, and multiple web mounts/unmounts. Measure additional download size, startup, interaction latency, retained memory and host-site style/router interference. Set performance budgets against the chosen host site before declaring success.
+The renderer spike should test one small interactive artifact in Dioxus Web and a desktop WebView. Cover stable focus and draft preservation, IME input, keyboard navigation, accessible names, stale events, restart, repeated revision swaps, and multiple web mounts/unmounts. Measure additional download size, startup, interaction latency, retained memory and host-site style/router interference. Set performance budgets against the chosen host site before declaring success.
 
 For marketing embeds, useful HTML must exist before the interactive renderer loads. Use a static/SSR projection from an exact revision and state snapshot, then attach interaction without changing that initial tree. Dioxus supports SSR/hydration, but its app-oriented approach does not itself prove lightweight island integration. Hydration, teardown and cache invalidation need explicit tests. A custom element can provide an integration wrapper; it is not a security sandbox. [Web mounting API](https://docs.rs/dioxus-web/latest/dioxus_web/struct.Config.html), [Dioxus SSR](https://dioxuslabs.com/learn/0.7/essentials/fullstack/ssr/), [custom elements](https://html.spec.whatwg.org/multipage/custom-elements.html)
 

examples/malleable-site/README.md

@@ -7,5 +7,5 @@ root to begin. The live browser demonstration is at
 [algal.computer/living](https://algal.computer/living/).
 
 An earlier renderer experiment consumed the exact same semantic view in Dioxus
-web/desktop and Ratatui. The spike concluded and the adapters were removed; the
-git history retains the record.
+web and desktop. The spike concluded and its adapters were removed; the git
+history retains the record.