Re: feedback requested on WAI CG Consensus Resolutions on Text alternatives in HTML 5 document

On Mon, 17 Aug 2009 12:36:19 +0200, Smylers <> wrote:

> Benjamin Hawkes-Lewis writes:
>> On 17/08/2009 10:30, Smylers wrote:
>> > If every instance of alt="" should also have role=presentation then
>> > could we get user-agents to infer that?
>> I'm suggesting that not every instance of alt="" should have
>> role="presentation"

(I agree)

> That seems reasonable -- and would be a good reason for not warning all
> authors using alt="" with a specific recommendation to add
> role=presentation.

Not necessarily. The logic runs like this:

There is lots of alt="" out there. Some of it refers to stuff that doesn't  
need more text for understanding, and can be effectively treated as  
presentation. Adding role="presentation" would help distinguish this, IF  
there is some mechanism where authors who care add role="presentation",  
while authors who just want to pass a conformance test don't.

In either case, I agree with Maciej that the spec should not tell software  
what not to show - dealing with these situations is messy enough that  
different heuristics for different users are going to matter for some time  
(at least until authoring improves, and I don't think that will be soon...)



Charles McCathieNevile  Opera Software, Standards Group
     je parle français -- hablo español -- jeg lærer norsk       Try Opera:

Received on Saturday, 22 August 2009 11:46:51 UTC