W3C home > Mailing lists > Public > wai-xtech@w3.org > August 2007

Re: non-conforming authoring tools: the backplane of portals, networking & photo sites

From: Lachlan Hunt <lachlan.hunt@lachy.id.au>
Date: Fri, 24 Aug 2007 19:46:24 +1000
Message-ID: <46CEA8F0.50302@lachy.id.au>
To: "Gregory J. Rosmaita" <oedipus@hicom.net>
CC: public-html@w3.org, wai-xtech@w3.org

Gregory J. Rosmaita wrote:
> the point of my pointing the WG to my.opera photo album, was NOT
> to display the world's most accessible photo portal site, but to 
> underscore 2 points:
> 1) it is not as difficult to describe complex images, as it has often 
> been argued on this list and cited as a reason why alt should no longer be 
> required nor longdescriptions except in certain cases
> i pointed james graham to http://my.opera.com/oedipus/albums not 
> to illustrate an accessible interface, but to counter the argument 
> that providing long descriptions is beyond the ken of most users,
> and to bolster my contention that a moment of reflection results 
> in some pretty perceptive perspectives on each photograph -- it 
> is the tool's fault that it reuses the tags associated with the 
> photo as pseudo-alt text, not the user's [...]
> so, PLEASE stop using my.opera album as a whipping boy for not 
> practicing what i and others preach

I didn't include your photo album in the wiki as an example of bad alt 
text, since that's clearly out of your control.  I included it because 
the alternative content provided by users in the comments is so good. 
It's evidence that users are capable of writing good descriptions, 
despite the unfortunate fact that it doesn't use alt or longdesc 
appropriately.  I certainly didn't intend to use it as a whipping boy.

> if anything, it is a compelling argument for stronger authoring tool
> conformance standards, NOT a reason do depracate a long description
> mechanism...

Exactly!  Any evidence that supports the addition of mechanisms for 
richer alternative content should be given some focus.

> 2) the problem with all photo album sites, YouTube and the rest lies
> not with the author, but with the underlying tools which prevent the 
> association of ALT or LONGDESC (or their successors) 
> any such tool MUST be defined by HTML5 as non-conforming authoring 
> tools...

The spec currently strongly encourages alt text to be provided, so 
perhaps it would be reasonable to require authoring tools to at least 
provide the ability for users to add it.  Although it would still need 
to define what to do in cases where users still don't provide it, which 
is what it currently does.

> the strategy cited in CSS2.1 for providing a long description is merely 
> the adaptation of a kludge that was deprecated when WCAG 1.0 became
> a technical recommendation -- use LONGDESC instead of an ambiguous
> "D" link, and have it programmatically bound to the object it describes 
> via markup, as we are dealing with content, not presentation, and the 
> CSS2.1 model is not only a step backwards, but leaves CSS incapable 
> user agents and their users nowhere

Note that the examples I included weren't included because they 
technically superior solutions, but that they were examples that 
demonstrated use cases for providing longer descriptions.

> the use cases should come first -- the page, after all, is dedicated to 
> mechanisms capable of providing a brief textual description of a 
> binary object and mechanisms capable of providing a longer, more 
> comprehensive description -- what was added by lachlan hunt to 
> the front matter belongs not under Use Cases -- which should 
> remain the first item on the page so as to frame the discussion --
> but in the "Research" "Further Reasearch" and "Suggested Solutions" 
> sections...

The problem with that page is that a list of people who benefit from 
from alternative text is not a list of use cases.  A description of what 
such people are trying to achieve in various cases would be.

A use case should describe the various types of situations that some 
feature would be useful.  For example, diagrams, flowcharts, 
photographs, screenshots, icons, etc.  For some of those, alt="" will be 
sufficient.  For others, a longer and richer alternative may be 
necessary.  By identifying those various cases, providing examples of 
them, and looking at them from both the authoring and end user 
perspectives in terms of what they are trying to achieve, the various 
solutions can be evaluated.

Maciej gave some good examples of what a use case should look like in 
IRC yesterday.  From the end user perspective:

"A user who can't make use of images for whatever reason visits a page 
that includes an image. She wants to understand the contents of the 
page, and the contents of the image are essential, but alt text is 
unlikely to be sufficient to convey the needed meaning."


And from the author perspective:

<mjs> "An author wants to publish a technical document including a 
flowchart that is essential to understanding the document. He wants his 
page to be accessible to users who can't make use of images, but the 
kind of short summary usually found in an alt attribute would not be 


Lachlan Hunt
Received on Friday, 24 August 2007 09:46:52 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:25:17 UTC