Re: Specifying longdesc to make it more useful, robust, and encourage better user agent implementation

Hi Josh,

>> Could you live with
>> or
>> or some combination of those two?

> I really like the second example, as an implementation of @longdesc.

I like pictures too. Another one could be the new tellmemore longdesc
extension from Opera [1]. It respects visual design yet can provide
critical functionality. It offers the best of everything.

Implementation screenshots are probably best handled as part of UAAG's
docs, such as their Implementation Guidelines. We could give the pics
to the User Agent Working Group and link to them there. But I am open
on that.

> Thanks to Stefan for his feedback.

Yes. Great catches, Stefan.

> and for pinging Steve about some editorial changes.

My pleasure, Josh.

> However, as a baseline (sic) I think we have more to gain from
> @longdesc being in the spec, than not.

Excellent observation. Baseline and technical details are what I have
been asking Jonas, Henri, and Ian about.

It would be good to know what they and other members of the working
group could live with in terms of longdesc spec text. In light of the
new research, identified use cases, and rationale [2][3] people may
have a new perspective.

If they do and if we can put our heads together to work out the
details maybe we can settle this amicably before Last Call as Paul

I've asked Ian if he would think about the matter [4].

We still haven't heard from Jonas or Henri. Jonas and Henri what are
your thoughts on:


Best Regards,


Received on Wednesday, 23 March 2011 17:19:44 UTC