Re: Change proposal for ISSUE-85

On Jun 15, 2010, at 10:46 PM, Jonas Sicking wrote:

> One thing that would be good to get clarified is how a UA is supposed
> to handle <a role=button>.

I believe this is already clear. The current WAI-ARIA Working Draft says:

"Host languages MAY document features that should not be overridden with WAI-ARIA (these are called "strong native semantics"). These can be features that have implicit WAI-ARIA semantics, as well as features where the processing would be uncertain if the semantics were changed with WAI-ARIA. Conformance checkers SHOULD signal an error or warning when a WAI-ARIA role is used on features with strong native semantics, even though user agents process the feature as described above."

<http://www.w3.org/TR/wai-aria/host_languages#host_general_conflict>

HTML5 defines <a> as having strong native semantics with an implied role of "link": <http://dev.w3.org/html5/spec/Overview.html#annotations-for-assistive-technology-products-aria>

HTML5 does not have any requirements that would override or contradict the WAI-ARIA requirements quoted above.

Thus, UAs are required to respect the role on <a role=button>, regardless of whether it is conforming or not.

> I think a good argument could be made that if someone has gone through
> the trouble to add 'role=button' to an element, then that is likely
> the most accurate role. Thus I think a reasonable argument could be
> made that we should forward this information to AT users.

I think you are right, and I believe this is already required by the combination of the relevant specifications.

Regards,
Maciej

Received on Wednesday, 16 June 2010 06:11:26 UTC