The architecture pages document the render/publish/mount pipeline and subtree
reuse thoroughly, but there's no developer-facing "why is my tap heavy / what
do I do" page. Add digging-deeper/performance.md covering:
- what a tap costs (event → render → publish → diff+paint) and that reactivity
is per-screen (child components are composition, not update boundaries);
- that native diff + paint scale with on-screen node count even when little
changed — the practical lever;
- levers in order: key lists, flatten the tree, split heavy screens into their
own route, use child components for reuse (not isolation), keep polling lean;
- how to measure with the built-in BenchmarkComponent (event/diff/paint
breakdown), with the real-device caveat.
Draft — opening early so it can grow in small, reviewable commits as I keep building a real NativePHP Mobile v4 app and hit things the docs don't yet cover. Each item below lands as its own commit; happy to split into separate PRs if maintainers prefer.
Included so far
Planned (separate commits, as I hit/confirm them)
Notes
🤖 Generated with Claude Code