Role parity for html:ins and html:del

[cookiecrook]

ins https://www.w3.org/TR/html51/edits.html#the-ins-element
del https://www.w3.org/TR/html51/edits.html#the-del-element

[aleventhal]

Good names for this are probably role="insertion" and "deletion". Would use this with ARIA annotations. I'm planning to implement an experimental version of this behind a flag in Chrome, meaning that it's an experiment that users have to turn on in chrome://flags.

[LJWatson]

Could I suggest "inserted" and "deleted" instead? Knowing that screen readers usually take the name of the role as the thing they announce, it would make more sense to hear something had been deleted than it would to hear it was a deletion.

[aleventhal]

@LJWatson it's not necessarily past tense. Sometimes it's a proposed insertion or deletion. The proposal is that can live inside a suggestion (future possible change) or revision (past change) container. The aria-details on the container points to an element that describes the type of annotation.

<span aria-details="suggestion-1">
  <span role="insertion">cat</span>
  <span role="deletion">dog</span>
</span>

where #suggestion-1 is something like:

<div role="annotation-suggestion">
Proposed change by Fru Fru Kittay
</div>

More info: https://github.com/w3c/aria/wiki/ARIA-annotations-draft-proposal

[asurkov]

Good names for this are probably role="insertion" and "deletion". Would use this with ARIA annotations. I'm planning to implement an experimental version of this behind a flag in Chrome, meaning that it's an experiment that users have to turn on in chrome://flags.

I would probably stick with ins/del naming to stay closer to HTML, less chances to confuse when using both HTML/ARIA.

Btw, why ARIA needs these roles upfront? Do HTML:ins/del elements have restrictions that ARIA allows to bypass or it's just part of some bigger effort to make ARIA embrace all semantics of HTML?

[aleventhal]

ARIAWG decided to go with the full name in order be consistent across ARIA with using full names. The only exception now is "img".

We need these roles as part of bigger effort to embrace semantics of HTML. Specifically, this allows an AOM tree to contain an insertion or deletion.

[cookiecrook]

@aleventhal wrote:

The only exception now is "img".

…and that mistake will annoy—nay—haunt one of the former editors until the day he dies. 🙄