W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > July to September 2004

Re: Long description and image on same page (was RE: [techs] Summary of techniques teleconference 22 September 2004)

From: Gez Lemon <gl@juicystudio.com>
Date: Fri, 24 Sep 2004 00:15:23 +0100
Message-ID: <009001c4a1c3$34328fc0$1900a8c0@juicy.com>
To: <w3c-wai-gl@w3.org>
MessageHi John,

John Slatin wrote:
> In a case where the conventions of writing in a field like art history
> or media studies, etc., make it appropriate for images and their
> descriptions to appear on the same page, it seems like we wouldn't
> want to require a *redundant* description via longdesc or a 
> description-link. But we might then want a way to indicate 
> programmatically that a certain portion of the text on the page is 
> explicitly intended to describe a particular image.

>Would the following work?
<p><img src="ohno.gif" alt="cartoon"></p>
<p class="caption">Figure 1. Cartoon showing members of WCAG WG with heads in their hands while Slatin harangues them.</p>
<p>Figure 1 shows a group of people around a conference table. They have buried their faces in their hands, except for one individual who is wearing a name tag that says "Slatin." Slatin's mouth is open........ </p>

Other than the context of the words, which could be argued is enough context, I don't see how this infers a relationship between the two paragraphs and the image. The solution I tend to favour, which turns out not to work with at least two screen readers, is to use the longdesc to link to the description on the page. The other idea that I thought might give some kind of relationship is to use a definition list, where the image is the term, and the description provides the relationship between the image and its text description. I've only thought this idea was good when there is more than one image that would require associated text.

  <dt><img src="..." alt="..."></dt>
    Associated text ...
  <dt><img src="..." alt="..."/></dt>
    Associated text ...

Even in that situation, I've gone on to provide a longdesc:

  <dt><img src="..." alt="..." longdesc="#snippet1"></dt>
  <dd id="snippet1">
    Associated text ...
  <dt><img src="..." alt="..." longdesc="'snippet2"></dt>
  <dd id="snippet2">
    Associated text ...

Is everyone in agreement that JAWS and Window Eyes should ignore the longdesc if the URI points to a fragment identifier on the same page? I've looked through the specs, and the association looks correct to me. I must be missing something somewhere. 

Best regards,


Supplement your vitamins 
Web: http://juicystudio.com
Keeping developers informed!
IWA/HWG Member
Received on Thursday, 23 September 2004 23:12:41 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 16 January 2018 15:33:51 UTC