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

Who needs what Re: A Call to Reorganize WCAG 2.0

From: Charles McCathieNevile <charles@sidar.org>
Date: Tue, 24 Aug 2004 15:44:08 +0300
To: "RUST Randal" <RRust@covansys.com>, w3c-wai-ig@w3.org
Message-ID: <opsc8ofuk3w5l938@widsith.local>

On Tue, 24 Aug 2004 07:33:27 -0400, RUST Randal <RRust@COVANSYS.com> wrote:

> Accessibility can certainly be automated, if it approached from a
> semantic standpoint. For example, lists of links should be marked up as
> unordered lists. Certain elements should have ALT and TITLE attributes
> attached.

deciding when a title is required on an image isn't something that can be  
automated. Any half-decent tool (few and far between, unfortunately) will  
be able to check if this image has been used in the same way before.  
Deciding if an image is decorative or not is virtually impossible to do  
automatically. The totally superfluous images that make Karl Dubost's  
presentations look cooler than mine are the same (and as someone with a  
clue about Web Architecture he uses them with teh same URI) as images in  
pages taht show where he has been.

Although there are ways of guessing that some text should be a heading  
(especially if it was styled consistently, getting it right through  
automation is a matter of pure luck.

> Perhaps accessibility validation should encompass the whole of the web
> page. What I mean by this is that a designer or developer should follow
> this process:
> 1. Validate HTML
> 2. Validate CSS
> 3. Validate Accessibility
> 4. Measure Usability

(How do you validate accessibility if you can only measure usability?)

> In step 3, we are testing the structure of the Web page, which fits in
> nicely with the rest of the W3C technologies.

I trust, then, that you have a tools which will take the instructions to  
Validate HTML, then check the CSS, before testing Accessibility and  
Usability, and recognise that it's an ordered list of instructions that  
can be readily illustrated with little diagrams that you can get from the  
web? Because I think that at least recognising it as a list it something I  
could rely on any person I would employ to do.

> Most designers actually test their pages in a browser without style
> sheet support in order to check accessibility. If the page displays
> logically and works within Opera without CSS and Javascript, it's going
> to be a very accessible web page.

That seems like a bold claim.

> Testing for accessibility needs to be as valuable, and as easy, for the
> designer/developer as validating HTML and CSS.

Here is where we fundamentally disagree. In my opinon, testing for  
accessibility needs to be as simple as possible, so long as it results in  
useful and accurate information. (I note that validating code according to  
syntax is not sufficient to determine whether you met requirements of the  
sepcification. Layout tables are not valid HTML 4 - it says so in the  
spec, and nor is using blockquote for formatting. But you can do both  
these things in Dreamweaver, and tools will tell you that you have valid  

I think what is most important is that testing for accessibility be a  
process that is clear enough taht developers can make content accessible.  
But then, I don't subscribe to "WCAG are writing things taht will be a  
law" - the WCAG group strie me as thoroughly unqualified to write law.  
hopefully they can explain accessibility - how to identify problems, how  
to do things right - in a way that is clear enough that lawyers and  
developers can make use of their work. In particular, smart tool  
developers can help by developing interfaces for authors that mean they  
only need to answer "common-sense" questions. To harp on an example "Does  
this outline cover the main points of the page?" is badly expressed, but  
more useful to the average person than "have you used headings  
appropriately with a correct nesting structure?".

just my morning 2 cents worth...


Charles McCathieNevile         charles@sidar.org
FundaciĆ³n Sidar             http://www.sidar.org
Received on Tuesday, 24 August 2004 13:44:44 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 13 October 2015 16:21:29 UTC