Is the documentation regarding the computation of name and description of tabular elements correct?

[matt-curtis]

According to the documentation for the tr, td, and th elements, the current computation rules only consider the aria-label, aria-labelledby, or title attributes, otherwise these elements have no name or description.

Should not the subtree of these elements be considered, at least as far as the accessible name goes?

[mmeller75]

Curious why this was closed. I was going to ask the same question. My understanding of the accessible name computation, which corresponds to how I've seen accessible names calculated by browsers, is that content should fall between ARIA labels and title. Looking at equivalent ARIA roles, they all accept name from content.

[scottaohara]

These elements should probably be broken apart and their naming re-reviewed.

for example:

• th should better communicate how abbr is used for naming
• td and th should likely allow name from subtree
• firefox issue about tr name - tr should not be named if a standard table (not grid)

[scottaohara]

from @extra808

Currently the name computation for tr, td, and th elements are the same. I don't have an opinion about the others but for th elements with the columnheader or rowheader role, shouldn't "Otherwise use th element subtree." be included, before "Otherwise use title attribute."?

https://w3c.github.io/html-aam/#tr-td-th-elements-accessible-name-computation

In practice, when a th columnheader contains text and has a title attribute, Chrome (v112) and Firefox (v111) do use the subtree for the name, according to their DevTools. However Safari's (v16.4.1 for Mac) DevTools show the cell's name as being the title attribute's value. Safari's name seems wrong but it's conforming to these mappings as currently written.