docs: zoom previews in the roadmap, gaps and control protocol (M8)

Refs #295

docs/control-protocol.md

@@ -93,7 +93,10 @@ ellipse, its arc handles `arcHandles`, for a selected rectangle, frame, polygon
 an idle window draws none) with the most frequent repaint requests behind them (`repaintCauses`:
 `[{at: "file:line", count}]`, every request made in the frame before each one, delayed ones such as
 the autosave deadline included; egui's `context.rs` begin-pass line means input events) and what the last canvas render had to do (`canvas.renderWork`: `full`, `partial` with
-the `pixels` re-rendered after a pan or an edit, or `reused`). While a prototype plays, `presenter`
+the `pixels` re-rendered after a pan or an edit, or `reused`), whether the canvas shows a coarse
+zoom preview or the refined image (`canvas.quality`: `preview` or `refined`) and the time from
+asking for a render to the frame showing its preview and its refined image (`canvas.previewLatencyMs`,
+`canvas.refineLatencyMs`; `null` until there was one). While a prototype plays, `presenter`
 (full window) and `preview` (the preview pane) hold the player's state: the `current` frame and its
 `name`, `start`, `history`, open `overlays`, how many delayed mouse triggers are `waiting`, the screen's `scroll` offset and `scrollRange`, the
 `device` (and whether it is turned to `landscape`, its `deviceColor`) and the `viewport` of the frame

docs/gaps.md

@@ -193,7 +193,9 @@ No alpha blockers: the alpha gate in [`roadmap.md`](roadmap.md#alpha-gate) passe
 
 ### 7. CPU canvas where Figma uses the GPU
 - **Missing:** GPU rendering of the canvas. FigCraft renders with tiny-skia on the CPU (banded,
-  parallel, retained) and shows the image through wgpu.
+  parallel, retained) and shows the image through wgpu. A new zoom can't reuse the retained
+  image; when renders from scratch take a frame or more, the whole view is previewed with a
+  sixteenth of the pixels first, then refined.
 - **Evidence (measured, `cargo xtask bench`):** a 50k-layer page ~8 ms zoomed out with levels of
   detail, single-layer edits at 10k layers ~0.2 ms; full 1440×900 renders 7–15 ms at 10k layers.
   Heavy effects (glass, background blur, noise) and 4K/5K canvases at 2x are where CPU time grows.