Re: Flickr and alt

On Tue, 19 Aug 2008 22:56:27 +0200, John Foliot <> wrote:
> Anne van Kesteren wrote:
>> Because we might be able to suggest something more reasonable
> I for one have been awaiting this suggestion for some time now.  Nothing  
> has emerged from the WG to suggest any real alternative.  Once a viable
> suggestion emerges, then the discussion can continue in that vein.   
> Please Anne, suggest something more reasonable, or table this line of  
> debate.

The proposal currently in the draft seems better than making the page  

>> than
>> making it compliant by just putting in some random string of
>> characters. If some photosite did care about validation for one
>> reason or another you might end up with alt="" everywhere, even on
>> the significant photos.
> According to the current draft, that would be non-conformant already, so
> what are you suggesting exactly?

It would be non-conforming, but would pass the validator, which I believe  
is the crux of the issue (for both sides).

>> This would preclude the photo from being
>> detected and people who can't see the photo won't be able to pass it
>> on to someone who does.
> Anne, this was the outcome for when @alt was being considered optional
> "sometimes"...
> ("I am currently following HTML5 (omitting alt) as it wasn't really  
> clear to
> me what would be a better solution given the single constrain I have: not
> finding it necessary to provide replacement text for all those images." -
> I am glad to see you have now come to understand at least one of the
> issues...

Actually, per the proposal at that time the image would be detected.  
However, popular implementation disagreed so that's probably why Hixie  
revised the proposal.

>> (Unless of course they use special tactics
>> for that site to discover the photos, but that's not really improving
>> the status quo I think.)
> Which shoots down the WHAT WG's earlier thoughts regarding "heuristic
> analysis" of images without alt text...

With "special tactics" I meant enabling specific features allowing them to  
detect images even though they have alt set to the empty string. Not image  
heuristics which hopefully going forward become better as people uploading  
photos are certainly not going to do it...

>> In other words, compliant content is not accessible per se, so trying
>> to test accessibility on the compliance level seems like the wrong
>> thing to do. It feels similar to all those people validating as XHTML
>> Transitional happily using <font> and <table> for layout without
>> having a clue as to what's going on.
> Agreed.  Without a real cost to non-conformance, it is a moot point.  Yet
> another reason to insist that all images must contain @alt - those that
> choose to ignore this will have non-conformant content, but so what?

It doesn't improve the status quo and might encourage people to enter alt  
text that makes it harder to figure out what is going on.

> Now if HTML5 were to be brave enough to not render images that lacked  
> @alt (or another directly linked text alternative to visual content) then
> conformance in this area would mean something.  It could be used to  
> teach, to inform, to ask that content authors raise the bar.  Those that  
> choose not to can always revert to HTML 4.01, which, while insisting on  
> @alt does
> nothing if/when that conformance requirement is omitted.  There is
> precedence in this with the way that Internet Explorer dealt with quirks  
> and standard html as declared via the DTD, and funny enough content  
> developers learned that lesson quickly and added that knowledge to their  
> daily
> workflow.

This is not a realistic proposal.

> There is no debate that much of this comes down to pride of ownership,  
> pride of development, and simply wanting the best.  Striving for these  
> goals is
> important, desirable, and a Good Thing.  But weakening requirements so  
> that content owners and creators can have a "badge" of conformance does  
> nobody
> any real good.

I do not see it as weakening requirements, I see it as addressing a use  
case that was previously not catered for.

Anne van Kesteren

Received on Wednesday, 20 August 2008 15:13:30 UTC