W3C home > Mailing lists > Public > public-html@w3.org > October 2009

Re: ISSUE-30 (Longdesc) Change Proposal

From: Joshue O Connor <joshue.oconnor@cfit.ie>
Date: Fri, 30 Oct 2009 16:01:03 +0000
Message-ID: <4AEB0DBF.6040902@cfit.ie>
To: Jonas Sicking <jonas@sicking.cc>
CC: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>, John Foliot <jfoliot@stanford.edu>, Charles McCathieNevile <chaals@opera.com>, public-html@w3.org
Jonas Sicking wrote:
> On Fri, Oct 30, 2009 at 7:02 AM, Joshue O Connor <joshue.oconnor@cfit.ie> wrote:
>> Leif Halvard Silli wrote:
>>> I agree: That @longdesc is a link is the key. @longdesc has other
>>> advantages as well, but they are all linked (sic!) to the fact that it
>>> is a link.
>>>
>>> I wonder if it will be possible to make aria-describedby into a link,
>>> though. [...]
>> Yes, to give @aria-describedby the ability to reference a URI as well as
>> a IDREF would be a great advantage.
> 
> Why not simply say that if the IDREF points to an element that is a
> link, then that link provides the description.

Yes. I would like to see As like screen readers being able to buffer
those descriptions (kinda like an AT) lookahead. I don't know if this
will work, or be best practice etc. Just an AT sci-fi, nice to have.

> This forces authors to consider putting the link to the description in
> the normal flow of the page, thus making it accessible to everyone,
> not just users of AT tools. Something which usually is a win.

Yes, and a good idea, but sometimes longer descriptors are needed where
the image may already contain some details that can be parse by a
sighted person but there is no textual description present, or to do so
would imply some kind of duplication. So to point to an external URI
would be a way around this I guess.

> For the cases when rendering the link is not desired, the link can
> simply be hidden using CSS or the @hidden attribute. Or you can place
> the link in the <head> which is never rendered.

Again, UAs that would need to find this link, may not be able to do so
currently. That could be changed as new behaviour could be modeled
depending on context. Again, this is partially a user agent issue.

Cheers

Josh
Received on Friday, 30 October 2009 16:01:55 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:51 GMT