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

Re: DRAFT response Re[3]: Request for PFWG WAI review of Omitting alt Attribute for Critical Content

From: Charles McCathieNevile <chaals@opera.com>
Date: Thu, 29 Nov 2007 12:58:12 +0100
To: joshue.oconnor@cfit.ie, "Al Gilman" <Alfred.S.Gilman@ieee.org>
Cc: wai-xtech@w3.org
Message-ID: <op.t2jx7aq4wxe0ny@widsith.local>

On Thu, 29 Nov 2007 11:59:54 +0100, Joshue O Connor  
<joshue.oconnor@cfit.ie> wrote:

> As an aside, while, in principle, I agree that there may be situations
> where there just is no suitable alternative textual description
> possible, and the main content whether an image a video object or
> whatever cannot be described.

I am at a loss to imagine such a situation.

> Then IMO using a alt="" null is fine.

In various cases where an object is described in the rest of the document  
it appears in, then indeed there are cases for having alt=""

The flickr examples so commonly given may well be something effectively  
like

<html xmlns="http://www.w3.org/1999/html">
<head>
   <title>Flickr Photo from Chaals</title>
</head><body>
   <h1>Me and my cat playing</h1>
   <p><img alt="" title="photo" src="FYDJ.jpg"/></p>
   <p>tags:</p>
   <ul>
    <li>chaals</li>
    <li>fanstasy</li>
    <li>mog</li>
   </ul>
  </body>
<html>

(for extra points, put an id on the header and make it the longdesc. For  
more points, have a real paragraph of description and make that the  
longdesc).

This provides the relevant available information (poor, but the kind of  
level we see typically). Repeating some of it as the alt is not all that  
helpful. In a more complete example, say a movie site, you might have

<html xmlns="http://www.w3.org/1999/html">
<head>
   <title>The Mog - Chaals' Movies</title>
</head><body>
   <h1>Playing with the cat</h1>
   <p><img alt="" title="movie in SVG" src="MogMovie.svg"  
longdesc="http://movies.example.org/TheFamousChaalsAndMog"/></p>
   <p>tags:</p>
   <ul>
    <li>chaals</li>
    <li>fanstasy</li>
    <li>mog</li>
   </ul>
  </body>
<html>

The kind of authoring interface that is relevant is one that asks the  
author if there is a description already available (URI or click on the  
description in the page), and if the image is part of the content  
(checkbox).

Where there is not enough information available to replace the image, such  
as a key graph presented in a report, then having no alt is better than  
alt="" - although it fails the testable validity requirement of having  
alt, this is a warning that there is a more serious problem (which there  
is). See ATAG checkpoint 3.4 [1] and associated techniques information [2].

Other relevant information in this (very old) discussion is available in  
the ATAG implementation report [3] (This was before W3C had teh CR  
process, but is a prototype CR report), the Working draft of the  
equivalent techniques for the draft ATAG 2.0 [4] (see B.2.4 and B.2.5  
which follows), inter alia.

[1] http://www.w3.org/TR/WAI-AUTOOLS/atag10.html#check-no-default-alt
[2] http://www.w3.org/TR/ATAG10-TECHS/imp3#check-no-default-alt
[3] http://www.w3.org/1999/12/atag-implemented
[4] http://www.w3.org/TR/2007/ATAG20-TECHS/tech2.html#check-no-default-alt

Cheers

Chaals

> In
> short, I would rather see authors _continue_ to use alt instead of them
> getting the idea that it *may* sometimes be left out.  Even if this *is*
> the case. I think this is one cowpath worth paving while suitable better
> methods are developed, if only to avoid somehow subtly devaluing alt in
> the mind of the author by producing badly authored examples of these
> edge cases where alt may be omitted in the spec.



-- 
Charles McCathieNevile  Opera Software, Standards Group
     je parle français -- hablo español -- jeg lærer norsk
http://my.opera.com/chaals              Try the Kestrel - Opera 9.5 alpha
Received on Thursday, 29 November 2007 11:58:30 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 27 April 2012 13:15:44 GMT