T.J. Crowder wrote: > > Is ARIA's "label" concept what Dr. Goss is looking for in this > case? (Whether via `aria-label` or `aria-lablledby`.) Or > "description" (e.g., `aria-describedby`, which doesn't appear > to have an inline equivalent)?
Hi T.J. Based on what I've seen and heard, yes, I think that aria-describedby would be the best fit today. Ideally aria-describedby would point to text on screen/on the page, which might introduce negative artifacts in other ways (??? "Learning" content?). The text could also be placed off-screen and referenced by aria-describedby, but it should be noted that any text off-screen is conveyed to the AAPIs as flat (non-markedup) text - or string text - so there is a minor issue there if there is a need to supply richer text. (However, since he first proposed adding @alt to a span, and @alt text is also string text, it is likely not an issue here). I personally liked the idea however of expanding the ability to add @alt to other attributes, including the in-line <span> element, and I think that this idea is worth teasing out a bit more. Dr. Goss (I believe it was he) also asked how to move this forward, to "ask the question". The answer is to open a bug in the HTML5 bugzilla, and/or propose it as a new feature to HTML.next (send an email with that in the subject line to the same mailing list here). HTH JF
