Re: longdesc="" in HTML5

Hi Ian,

On Sep 7, 2008, at 11:36 AM, Ian Hickson wrote:

> I hate to throw fuel on the fire here, but as far as longdesc="" is
> concerned, the reason it isn't in HTML5 is that there has never been  
> any
> feedback sent that described a problem for which longdesc="" was even
> remotely considered as a solution.
>
> As is described here:
>
>   http://wiki.whatwg.org/wiki/FAQ#Is_there_a_process_for_adding_new_features_to_the_spec.3F
>
> ...and as has been previously described on this list, the way that any
> feature gets added is by first establishing a problem to solve, and  
> then
> coming up with the best solutions to address the problem. To get on my
> radar, there has to be an actual problem we are trying to solve that  
> isn't
> already solved in some better way. This has never happened for any  
> problem
> where longdesc="" (or things like it) were a candidate solution.

At this point Ian, you're plugging your ears and screaming "I can't  
hear you" when you say something like that. You need to stop acting  
like a child and step up and participate in this WG and be a real  
editor. There have been countless emails and wiki contributions[1]  
describing the need which long description fulfills. Let me take  
another stab at it here.

Distilling the many discussion on the topic by this WG, the longdesc  
attribute provides three primary functions as I see it (functions that  
should be supported natively by HTML). In part it was designed to  
enhance the alt text (text replacement) of the alt attribute since the  
IMG element provides insufficient support for alt text that is lengthy  
(more than a single paragraph) or richly marked up (with emphasis,  
headings, etc). However, because of the longdesc attribute's  
peculiarities (e.g., the attribute's name and the need often for a  
separate document to contain the referenced document fragment) it has  
come to serve a separate function: providing descriptive text  
equivalents.

So there are three separate and distinct needs that have been  
articulated within this WG:

  1) an 'aria-describedby' like explicit association between an  
embedded image and textual description of the image
  2) descriptive text equivalent (often independent of the document)  
and closely tied to the image itself and not alt replacement text
  3) a lengthy structured (markup rich) alt text replacement text  
(only needed because IMG is a void element)

Clearly, ARIA can satisfy the first function better than the longdesc  
attribute does as James Craig suggested[2]. For descriptive text  
equivalents — especially when they are independent of the media's  
embedding within the current document — could better be handled by the  
proposal to expose the immanent media metadata through DOM attributes,  
methods and UI[3]. However, the last role — of providing a mechanism  
for lengthy (even just more than one paragraph) and markup rich alt  
text is not addressed by these. It may be that we also need the burden  
of the other two functions lifted off of the longdesc attribute and  
provided in these other ways. However, as long as we are stuck with  
the img elements as a void element we need to provide a mechanism for  
alt text that cannot be served through the string contents of an  
attribute like the alt attribute.

When limited only to the third function, it is clear the longdesc  
attribute is not needed on any element except the img and the embed  
elements.

Some solutions for the alt text function of the longdesc attribute?

1) Leif raised the possibility of advising authors to use the object  
element for embedding images[4] in circumstances where more than one  
paragraph and/or other markup is needed. This should work well in  
current browsers (once IE8 is released at least), but for legacy  
browsers, authors need advice on what to do.

2) For the future, we could also make the src attribute a universal  
attribute to embed image replacement using any element. This could  
even be made to work with leading edge CSS3 implementations solely  
with a change to the default implementation style sheet so  
implementations needing to implement CSS3 would already get HTML5  
support for a universal src attribute.

3) Finally, at the very least we should make use of the contents of  
the IMG element in the XML serialization (taking precedence over the  
alt attribute if any). There may be other solutions, but these seem  
like the obvious ones to me.

I hope this helps. Of crucial importance for any WG, is an editor who  
understands and follows the deliberations of that WG.

Take care,
Rob

[1]: <http://esw.w3.org/topic/HTML/LongdescRetention>
[2]: <http://lists.w3.org/Archives/Public/public-html/2008Sep/0186.html>
[3]: <http://esw.w3.org/topic/HTML/UANormAndDOMForMediaPropeties>
[4]: <http://lists.w3.org/Archives/Public/public-html/2008Sep/0242.html>

Received on Monday, 8 September 2008 14:52:48 UTC