- From: David Bolter <david.bolter@gmail.com>
- Date: Tue, 12 Oct 2010 15:19:48 -0400
- To: James Craig <jcraig@apple.com>
- CC: Andi Snow-Weaver <andisnow@us.ibm.com>, Steven Faulkner <faulkner.steve@gmail.com>, W3C WAI-XTECH <wai-xtech@w3.org>, HTML Accessibility Task Force <public-html-a11y@w3.org>
Thanks James, On 12/10/10 3:02 PM, James Craig wrote: > On Oct 12, 2010, at 8:15 AM, David Bolter wrote: > >> The browser already knows what is visible and what is not. The reason >> aria-hidden was added in the first place was for assistive technology >> that looked directly at the DOM (and didn't rely solely on the >> accessibility tree exposed on the desktop). > > That's one of two reasons. > >> Personally, I would have preferred aria-hidden be removed from the >> spec than have it actually prune the accessibility tree. > > Many technologies have the potential for abuse, but there remain > plenty of legitimate cases where it is appropriate to hide a visible > element from the accessibility tree. We *have* to provide a way for > that to be made possible, or web developers will just avoid making > their apps accessible. We have role="presentation". > > Desktop accessibility APIs already allow this; now the web does, too. > >> What should we do if focus goes into this aria-hidden section of the DOM? > > The implementation guide covers that in: > http://www.w3.org/TR/wai-aria-implementation/#keyboard-focus_tabindex > > "All elements which are focusable MUST be exposed in the accessible > tree, so that when they receive focus, there is an object for which to > fire an accessibility event." > > Although it is somewhat vague as to who should be doing the work. It > should probably say: > > "User agents MUST expose all focusable elements in the accessible > tree, so that when they receive focus, there is an object for which > to fire an accessibility event." > Right, and this is a cardinal rule in gecko's accessibility implementation. I see ambiguity here though... should the focused element (inside an aria-hidden tree) be connected in the accessibility tree? I don't really want to explore this in time consuming detail, but just pointing out there is potential muck here. cheers, David
Received on Tuesday, 12 October 2010 19:20:27 UTC