- From: James Craig <jcraig@apple.com>
- Date: Tue, 11 Nov 2014 18:58:32 -0800
- To: Alice Boxhall <aboxhall@chromium.org>
- Cc: Steve Faulkner <faulkner.steve@gmail.com>, W3C WAI Protocols & Formats <public-pfwg@w3.org>
On Nov 10, 2014, at 2:14 PM, Alice Boxhall <aboxhall@chromium.org> wrote: > On Mon, Nov 10, 2014 at 1:12 PM, James Craig <jcraig@apple.com> wrote: > >> Quoting from the next line: "Because the element is no longer an image, the role-specific host language labeling mechanism (e.g. img@alt) no longer applies." > > I was actually wondering about that: why does the role affect the alternative text computation? The @src attribute will still work, after all. Why would we want to ignore the @alt attribute? Is there any chance the aria-label will not simply be a duplicate? The text alternative computation already takes role into consideration. For example, if will not traverse the contents of roles that do not list "roleFrom: contents" I could be convinced to allow @alt here, but if the user redefined the role, why should be role-specific labeling mechanism still work? If we allow @alt, should we allow other @summary on <table role="group" summary="foo">? Is that a slippery slope? What about labeling mechanisms in other host languages like SVG? James
Received on Wednesday, 12 November 2014 02:59:01 UTC