- From: Ian Hickson <ian@hixie.ch>
- Date: Wed, 21 May 2008 10:52:01 +0000 (UTC)
- To: Steven Faulkner <faulkner.steve@gmail.com>
- Cc: public-html@w3.org
On Wed, 21 May 2008, Steven Faulkner wrote: > > > > Note that the HTML5 spec as written today _does_ require the alt="" > > attribute when it is at all possible to include it. > > So does Anne's seminal HTML5 example of not including a text alternative > using the alt attribute [1], and reasoning [2] for not providing text > alternatives, result in conformant HTML5 or not? > > [1] http://anne.is.weggeweest.nl/image-viewer#2006,boston,9 > [2] http://annevankesteren.nl/2007/09/alt If it was possible for Anne to include alternative text, then it should be included. I can't tell whether this is a case of that or not because the image viewer in question doesn't work on the browser I tested it with (so there are much, much bigger accessibility problems for me on that site than whether the image has alt text or not). If an author doesn't provide alternative text because he doesn't care to, then the document isn't conforming. If an author doesn't provide alternative text because doing so wouldn't be practical, then the document is conforming, but violates a "should", and conformance checkers should probably give warnings to that effect. It is mostly up to the author to determine when the case goes from one to the other. In practice, providing useful alternative text for photographs is non-trivial and so such effort can quickly become impratical. That's just the way things are; making alternative text required without addressing this fundamental problem is an ostrich approach to solving accessibility. -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Wednesday, 21 May 2008 10:52:42 UTC