- From: Charles McCathie Nevile <chaals@yandex-team.ru>
- Date: Tue, 05 Mar 2013 00:34:26 +0100
- To: "WAI XTech" <wai-xtech@w3.org>, "Sailesh Panchang" <sailesh.panchang@deque.com>
- Cc: "HTML Accessibility Task Force" <public-html-a11y@w3.org>
On Mon, 04 Mar 2013 21:00:03 +0100, Sailesh Panchang <sailesh.panchang@deque.com> wrote: > As longdesc is meant to be exposed to all users and is meant to > function like a link, perhaps following points need to be clarified: > > i. How should user agents convey that the image has a longdesc attribute This is fundamentally an issue of user interface. Some screen readers announce there is a description available. My extensions used a button whose icon changed (the button also provided a guide to *which* images had longdescs, and let you quickly get to them), and then a button that appeared when there is a longdesc. Other approaches have added a marker to the image itself. I don't think we have sufficient experience to say that one or other kind of user experience is clearly better and should be adopted by everyone. But it is clear that no discovery is worse than users being able to discover the link. Which is why the spec says user agents should make it discoverable. > ii. Does it appear like a link and is it distinguished from other > hyperlinks on the page in any way? That depends on the user agent, as noted above. In practice I haven't seen any implementation that does this, and suspect that in general that would be a bad idea. > iii. The image will not be tab-navigable but the 'link' associated > with the image via longdesc will be tab-navigable and > keyboard-operable, right? This is part of being discoverable, and it is part of a user agent being useful. If it isn't, the implementation needs fixing, IMHO. > iv. What happens when tabindex attribute is used on focusable elements > on the page? Should the image with the longdesc also have a tabindex > attribute? If you're using tabindex to create a different order, then yes. Personally I think that's a bad idea in most circumstances, but I am sure there are cases where it makes sense. I suspect that in some minority of such cases, there is no disadvantage in leaving tabindex off a longdesc (e.g. because the most natural place for it to be in the tab order is last). But this is the sort of authoring practice that I expect to be developed into best practice recommendations by people like you, rather than things the specification should try to rigidly enforce, at least at the current stage of maturity... cheers Chaals > Thanks, > Sailesh Panchang > Tel 703-225-0380 ext 105 > www.deque.com > > keyboard user may not be -- Charles McCathie Nevile - Consultant (web standards) CTO Office, Yandex chaals@yandex-team.ru Find more at http://yandex.com
Received on Monday, 4 March 2013 23:35:02 UTC