W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > April to June 2005

Re: Re : Influence of valid code on screen readers

From: Maurizio Boscarol <maurizio@usabile.it>
Date: Sat, 18 Jun 2005 16:23:00 +0200
Message-ID: <42B42E44.8050003@usabile.it>
To: Matt May <mcmay@w3.org>
CC: Roberto Castaldo <r.castaldo@iol.it>, w3c-wai-gl@w3.org

Matt May wrote:

>> We're not in a position to say what is and is not "allowed". The WCAG 
>> WG is
>> not the Web police. 
>

I agree. I think that validation is important as general QA check, but 
it is overstated in regard of web accessibility.

In particular I think validation is not:
1. A preliminar accessibility requisite.
2. A vital check point to decide if a page is accessibile.

I indeed think validation is:
1. A quick and dirty check of quality of code (but not the only one)
2. Sometimes, but not always, a quite rough predictor of accessibility, 
especially in this stage of web evolution (people who care about code 
tend to care also about accessibility, but this may change in different 
periods; there isn't any guarantee);

I see validation rather like a "triple A" issue, than a "single A", just 
to speak in an old fashion way.

I'll try to explain why.

If validation were a basic accessibility issue, than the following 
should be true:
1. All invalid pages couldn't be accessible
2. A valid page should be always prererred to an invalid page

But I think both are false.

1. Can an invalid page be accessible? Well, yes. There are different 
example out there. If I have well structured contents, alternative 
equivalents, but some invalid characters or some invalid attribute, that 
may not be a problem for disable users - unless the user agent can't 
render the page: but with text/html mime type content user agent can 
render even invalid content, as they have always done, and so the page 
still remains accessible.
2. An invalid page could have a better code than an invalid one. I put 
an example online. This page ( <http://www.usabile.it/esempi/chilosa/> ) 
is valid, but i think it's not a good example of coding. There's another 
page here ( <http://www.usabile.it/esempi/nonvalido/> ), that isn't 
valid, but is far better coded, and far more accessibile due to its 
semantic/structural markup. This new page is invalid because some minor 
issue. The rendering is quite similar for the two pages.

The example is useful to highlight a confounded argument inside the 
"validation question". I think most people when talking about validation 
assume that they are talking about good structured and semantically rich 
pages, in opposition to old bad tag soup. This is often true, but not 
necessarily always.

Some valid pages could have very poor markup, as I have shown. There are 
a lot of news sites that don't uses tables for layout, but are filled by 
non-sensical div, they use div for almost everything, and this is not a 
very accessibile practice, despite of being valid or invalid! Structural 
and semantically rich code is very important, and this is not the same 
as validation! They are sometimes related, but not always.

Finally, there are also some pages that are invalid for some minor 
problems, but do have good accessibility. I recently delivered some 
templates in valid xhtml 1.0. When implemented in a CMS (by another 
staff), the pages were nomore valid. Why? Because they added some 
<script language="..."> tag and they didn't escaped a lot of ampersand 
in the urls!

So, the pages don't validate. But are the pages inaccessible? Damn'it, I 
don't think so. I think they are quite good pages, even if they are not 
valid.

In today's web many valid templates are made invalid by some bad 
implementation in old times CMS. I suppose it's a transitional effect. 
In tomorrow's web the new CMS will reach validation with little or no 
problems.

Still, that will not guarantee accessible pages. It depends on quality 
of code, of textual equivalents, of structural and semantic markup, and 
many other design solutions.

I think we should face with this new and hopefully temporary reality: 
the problem about validation isn't always tag soup, but some little 
implementation mistakes, that don't significatively affect the quality 
of code. The pages can nonetheless be accessible.

For all those reasons, I think validation is largely overstated. It's a 
good indicator of quality, but even minor problems in validity don't 
affect accessibility.

We should fight tag soup, and evangelize about good structural and 
semantically rich markup. But WCAG don't need to be mandatory on 
validation, at least not at the basic level.

I personally think the new italian law made a mistake including valid 
xhtml 1.0 strict (and only this dtd!!) for any new accessible pages to 
be made. I think the CMS market is not mature, and there's no reason to 
do such a big step forward right now. But this is an italian problem (we 
have bigger problem than this, actually, so we like to concentrate on 
the smallest things to feel better... ;-) ).

I think along with Matt May: wcag are not web police. We should not 
complain about validation, but only about real accessibility. Validation 
should be encouraged, but at the top quality level, not at basic level.

Maurizio Boscarol
Received on Saturday, 18 June 2005 14:14:16 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 23:39:37 UTC