- From: Maciej Stachowiak <mjs@apple.com>
- Date: Mon, 17 Aug 2009 03:24:10 -0700
- To: Steven Faulkner <faulkner.steve@gmail.com>
- Cc: Benjamin Hawkes-Lewis <bhawkeslewis@googlemail.com>, Ian Hickson <ian@hixie.ch>, HTMLWG WG <public-html@w3.org>, W3C WAI-XTECH <wai-xtech@w3.org>
Consolidating replies: On Aug 17, 2009, at 3:02 AM, Steven Faulkner wrote: > hi all, > > i think the following examples should raise a conformance error. > <a href="#"><img src="delete.png" role="presentation"></a> > <a href="#"><img src="delete.png"></a> > <a href="#"><img src="delete.png" alt=""></a> > <a href="#"><img src="delete.png" role="presentation" aria- > labelledby="this"></a> <span id="this">delete</span> > <a href="#"><img src="delete.png" alt="" aria-labelledby="this"></a> > <span id="this">delete</span> I believe HTML5 as currently drafted would make all of these conformance errors, because it requires non-empty alt when an image is the sole content of a link: <http://dev.w3.org/html5/spec/Overview.html#a-link-or-button-containing-nothing-but-the-image >. HTML5 makes non-empty alt mandatory in this case, and the requirement is machine-checkable. > I think the follwoing should not > <a href="#"><img src="delete.png" aria-labelledby="this"></a> <span > id="this">delete</span> I believe this is not currently conforming, but probably will be once ARIA is integrated. On Aug 17, 2009, at 3:15 AM, Steven Faulkner wrote: > hi henri, maciej > > >Wouldn't it be simpler to change a handful of browsers to hide <img > alt=""> from MSAA than to get zillions of authors to add > role=presentation where they >already have alt=""? > >Would it be appropriate, in light of this, to add a user agent > requirement that an img with empty alt should not be mapped to > accessibility APIs at all? It >seems like that would do the job a > lot faster than migrating content to role="presentation". > > I don't think the intent is to get all legacy content using > role="presentation" it is to promote the use of a standard generic > mechansim for elements that should not be exposed via the > accessibility API. the use on <img> being one example. OK, thanks for clarifying. In light of this, and Ben's feedback, I think it might be better not to have the warning. I think alt="" serves the purpose fine > > i think that it would be appropriate to add a user agent requirement > that an img with empty alt should not be mapped to accessibility > APIs as this is what it how it works in AT. Given Ben's feedback, it sounds like it may be better to map these images to accessibility APIs, and have a requirement that AT should not expose them unless it detects an error where it would be better to flag the existence of the image in some way. Thinking about the img-in- a-link example from Ben and how we could make it work in Safari +VoiceOver, I think we'd want to let VoiceOver flag the image in that case. > as far as the warning goes i can live with it not occuring. All right. Curious if anyone else has input. Regards, Maciej
Received on Monday, 17 August 2009 10:24:52 UTC