- From: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>
- Date: Tue, 8 Jun 2010 11:36:44 +0200
- To: Karl Dubost <karl+w3c@la-grange.net>
- Cc: Ian Hickson <ian@hixie.ch>, Bruce Lawson <brucel@opera.com>, HTML WG <public-html@w3.org>, HTML Accessibility Task Force <public-html-a11y@w3.org>
Karl Dubost, Mon, 7 Jun 2010 20:05:10 -0400: > The difference between the two examples illustrates two schools of thoughts. > > [ > cut two examples with > figure/img/figcaption > see for references > http://www.w3.org/mid/Pine.LNX.4.64.1006072239310.22659@ps20323.dreamhostps.com > ] > > > 1. Descriptive > In which the alt tag is used to described what is in the image. > > 2. Text flow > In which the alt tag is given an empty value (alt="") > because it would be awkward in the text flow without images. > > Both cases could be valid. Indeed. However, HTML5 has generally been a context where the "Text flow" school of thought (also known as "visible meta data") has ruled the ground. I am therefore (theoretically spoken) surprised to see Ian's lay-out of the problem ... > It depends on the context and what the text author would like to > convey to the readers. > It also depends on how "figcaption" element would be treated by > assistive technologies. > Something part of the text or something special with a marker. > > For example, in a vocal browser saying something like: > "IMAGE CAPTION blablabla" > or just > "blablabla" > > If the vocal browser says "IMAGE CAPTION" then the alt with a > meaningful content seems to be necessary. I think the problem is whether we see @role and aria as an integral part of HTML5 that user agents *must* support. Or if the view is that WAI-ARIA is like a "Cascading Accessibility Sheets", which only some are required to support. E.g. if we do this: <figure role='image'> <summary>blablabla</summary> <img src="i" alt="" role="presentation"> </figure> Then, firstly, a vocal browser *would* say "IMAGE CAPTION blablabla". The problem is those UAs that do not support WAI-ARIA. What should we do for them? I see these options: 1) Do nothing. After all, even if they don't support WAI-ARIA, they are required to support <figure>. 2) Do this: <img alt="See summary!" role="presentation" src="i"> (This could probably be confusing.) 3) Duplicate the figure caption inside the @alt. This wouldn't bug those who have UAs whic support WAI-ARIA. I wonder if not 1) is the best approach. Another approach could be to make sure that @alt and the the figure caption are used for different things (or at least contain different text). E.g. in Ian's example, the @alt contains optional information and thus that @alt falls into the aria-describedby category (while the figure caption, of course, falls into the aria-labelledby category): <figure role="img" aria-describedby="img-element" > <img id="img-element" src="ceo" alt="Photograph: A tall man with brown hair, and a moustache sits at a keyboard while smiling to the camera." > <figcaption> Brian Slick, CEO of Blammo Corp, leveraging a synergy. </figcaption> </figure> -- leif halvard silli
Received on Tuesday, 8 June 2010 09:37:51 UTC