search

cvs-health / testaro

A thousand web accessibility tests performed by 11 tools
MIT License
29 stars 10 forks source link

styleDiff test questioned #8

Closed jrpool closed 2 years ago

jrpool commented 2 years ago

A commentator believes that the styleDiff test treats benign style differences as defects, even when they enhance accessibility, so it should not be used in a battery of accessibility tests, unless it is modified to report only harmful style inconsistencies.

jrpool commented 2 years ago

The styleDiff test is limited to list links, adjacent links, buttons, and headings (of each level), and to a set of style properties enumerated in the tests/styleDiff.js file. It is a test of consistent identification of components with particular functionalities (WCAG Success Criterion 3.2.4). Variability in these style properties can be expected to make identification of components more difficult. The test quantifies variability, so a page with only a small set of styles for a particular component is distinguishable from a page with a large set of such styles. Therefore, the test seems reasonable. Any user who considers it unhelpful can avoid running it or avoid using it in scoring.

However, some changes in the particular style properties applied to particular component types are being made, in light of this comment.

A revised version of styleDiff is incorporated in version 4.10.0.