Benchmark UI interactions against the capacity budgets
- github
- github/openchronology/timeline-app
- commit
- main
- github/athanclark
bench/ui.mjs drives the editor with real input (Playwright mouse, keyboard and touch) and times 24 interactions, from the input event to the frame that shows the result, against budgets in src/capacity.ts derived from the advice thresholds: response 100 ms, view and gesture frames 150 ms, results 1 s, open 1.5 s. It checks both directions: within COMFORTABLE sizes every interaction meets its budget, and beyond them a miss must coincide with the in-app advice (otherwise an "advice gap"). Runs the in-memory browser editor (full speed and 4x throttled) and the platform (production Next.js with PostgreSQL, signed in). An injected probe does the timing without browser-specific APIs or eval, so the desktop WebKit can be driven the same way next. The report gains an Interactions table; CI gates 1,000 moments (1.5x tolerance); the Benchmarks workflow runs the full sizes. Refs #51. Findings filed as #68 (navigation speed) and #69 (advice and comfortable sizes). Co-Authored-By: Claude Opus 5.5 Claude-Session: https://claude.ai/code/session_01Ti36soiZpKdSxKfvQLEcfY https://github.com/openchronology/timeline-app/commit/481cbb9c8b7aceb36ade2e2218822ffdf003f607
Related
Open Benchmark UI interactions against the capacity budgets in the timeline