Deep Dive into Compose Semantics
summary
Live demos show how TalkBack composes announcements and concrete fixes: use contentDescription, role, mergeDescendants to label clickables and group text, use clearAndSetSemantics + customActions to turn complex feed cards into one focusable item, use stateDescription and liveRegion (polite vs assertive) for downloads, and recover TalkBack focus with a focus ID + wait-for-frame trick. Examples include animated word‑by‑word streaming fixed by clearAndSetSemantics and layout‑inspector checks; good if you need actionable Compose accessibility patterns.
AI-written summary · about
chapters (12)
- 0:00Introduction
- 1:02Semantics basics
- 2:47Unlabeled buttons
- 5:16Content description vs label
- 8:27Updating UI state
- 11:20Merging descendants
- 14:51Clear set semantics
- 16:38Simplifying complex content
- 20:19Streaming text animations
- 25:16Recovering focus
- 30:55Inaccessible UI patterns
- 33:26Summary
related talks
19mCreating Narratives: Shaping TalkBack with Compose Semantics
13mBuilding accessibility support for Compose
44mAccessibility with Compose : Experimentation & Live coding
43mBuilding Accessible Apps with Jetpack Compose
34mAutomating Full-Screen TalkBack for a Government App with Designer Collaboration and AI Agents — a Practical Experience
42mBuilding Compose Apps for everyone
27mLeverage Compose Tooling to Improve Your Workflow
23mComposing With Confidence
