REAL-WORLD INTERPRETIVE
Count what is simultaneously visible
Record markers, text labels, moving layers, event pace, and dossier length at the same viewport and task. Hidden layers and progressive disclosure should be documented rather than assumed.
REAL-WORLD INTERPRETIVE
Declare access parity
State whether source context, keyboard operation, reduced motion, and low-bandwidth alternatives are actually available. A promise to add them later does not satisfy the review.
REAL-WORLD INTERPRETIVE
Interpret pressure, not a user score
The workbench aggregates design pressure to prioritize review. It does not measure intelligence, attention, disability, diagnosis, competence, or an individual user's capability.
REAL-WORLD INTERPRETIVE
Use the remediation list
Clustering, label prioritization, event staging, progressive disclosure, static alternatives, semantic tables, and explicit source panels are suggested when the related trigger is present.
REAL-WORLD INTERPRETIVE
Follow with representative testing
Test comprehension, task success, errors, context recovery, workload, keyboard completion, screen-reader output, motion comfort, mobile use, and constrained bandwidth.
REAL-WORLD INTERPRETIVE
Define a stop rule
Do not ship a feature merely because it increases engagement. Confusion, source invisibility, loss of keyboard access, motion harm, or inaccessible fallback should trigger revision or rollback.