Identify Purpose
In content implemented using markup languages, the purpose of user interface components, icons, and regions can be programmatically determined.
In plain language
Icons, controls, and regions of a page should be marked in the code with what they are, so their meaning doesn't depend on a picture that only makes sense visually.
Why it matters
Some people with cognitive disabilities benefit from replacing icons with a symbol set they understand better, or from simplifying a page by hiding non-essential regions. Assistive tools can only do that if the code says what each icon or region is.
In practice
Fails this criterion
A page has navigation icons for home, search, and settings, but nothing in the code identifies what each icon represents. They're just images with no semantic meaning.
Meets this criterion
The same icons and page regions are marked with proper roles and semantic attributes, so assistive technology can identify them and, if the user wants, substitute or simplify them.
Test it yourself
Check the site's icons and major regions (navigation, main content, search) in browser dev tools. Are they marked with roles or attributes that describe their purpose, or are they unlabeled images and generic containers?