Accessibility · Cluster
Assistive Technology
Screen readers, switches, and input methods.
- 01 The Screen Reader Announcement Pipeline How a screen reader turns an accessibility-tree event into spoken output, from event dispatch through speech synthesis, and where latency and interruption come from.
- 02 Accessible Name Computation How a screen reader derives the single name string for an element from text, labels, aria-labelledby, and fallbacks, and why precedence rules matter.
- 03 Switch Access and Row-Column Scanning How a single switch lets a user navigate every control on screen through timed scanning, and how the scan timing and group layout determine entry speed.
- 04 Braille Displays and the Translation Layer How a refreshable braille display turns screen text into dot patterns through character encoding, and why contracted braille requires a separate translation step.
- 05 Text-to-Speech Synthesis Under the Hood How a speech engine turns an SSML document into audio through text normalization, phoneme conversion, and prosody, and where screen-reader latency comes from.
- 06 Live Regions and Asynchronous Announcements How aria-live regions cause screen readers to announce content that changes without focus moving, and how the polite and assertive politeness levels are implemented.
- 07 Virtual Cursors and Browse Mode How a screen reader maintains a separate reading cursor over a virtual buffer of the page, distinct from real keyboard focus, and why the two desynchronize.
- 08 Debounce, Dwell, and Key-Repeat Filtering How assistive input systems distinguish a deliberate activation from switch bounce, tremor, and held keys, and how dwell and repeat timing are computed.
- 09 Eye Tracking and Gaze-Based Selection How an eye tracker estimates where a user is looking, why gaze alone is a poor pointer, and how dwell and other schemes turn gaze into reliable selection.
- 10 Word Prediction and Accelerated Text Entry How word prediction and abbreviation expansion reduce the keystroke cost of text entry for switch and scanning users, and how the prediction model trades accuracy for input rate.