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

Re: Should validity be P1 or P2? (was RE: summary of resolutions from last 2 days)

From: Christophe Strobbe <christophe.strobbe@esat.kuleuven.be>
Date: Tue, 21 Jun 2005 16:50:49 +0200
Message-Id: <>
To: Matt May <mcmay@w3.org>, Joe Clark <joeclark@joeclark.org>
Cc: WAI-GL <w3c-wai-gl@w3.org>


At 19:39 20/06/2005, Matt May wrote:

>Joe Clark wrote:
>>Matt May wrote:
>>>You've completely missed the other half of the problem, which is that we 
>>>can find many examples of valid pages that are plenty inaccessible.
>>That is an unsupported claim. Nobody has provided even the standard three 
>>*real-world* examples that I repeatedly call for and never get.
>You don't get them because you're not the boss of me, or anyone else for 
>that matter, and nobody has agreed to your criteria but you.
>I have already outlined a number of ways in which valid code can be 
>inaccessible, which I recap here in handy list form:
>- Non-semantic HTML (<b> or <font> instead of <h*>, <br> and * instead of 

Example 1: the homepage of http://www.hyfinity.com/, which claims valid 
XHTML 1.0 (which is correct: 
http://validator.w3.org/check?uri=http%3A%2F%2Fwww.hyfinity.com%2F), valid 
CSS and WCAG AAA conformance. However, the XML declaration reads
<?xml version="4.0" encoding="UTF-8"?>
The body consists only of <div>, <span> and <a> elements; there is not a 
single <hx> or <p> in sight.

Example 2: the sitemap of the previous site 
Lists are faked with nested <div> elements and CSS. Screen readers can 
access the text, but the user does not know where the site navigation (at 
the top of the page) ends and where the actual site map starts.

By the way, one of the "solutions" provided by this company is "WAI 
Compliance" (http://www.hyfinity.com/WAI%20Compliance.html), but their site 
is a rich resource of non-semantic markup. Three cheers for Hyfinity!

Example 3: http://gallus.forestry.uga.edu/ggg/protocols/. Valid XHTML 1.0 
Strict (see 
This page is on "Protocols", but this text is in a <div> and an <a> instead 
of an <hx>; subsection titles are in <h1> but should be in a lower level 
heading if "Protocols" is marked up as <h1>. The "About" page 
(http://gallus.forestry.uga.edu/ggg/about/) is a similar case.

Example 4: http://www.brainmass.com/content/about/. (Almost valid XHTML 1.0 
Transitional: the validator reports: "non SGML character number 146"; see 
Here, the title of the page ("About BrainMass") is faked with an image with 
empty alt text; the image is not in a <hx> element but in a layout table. 
(Section titles use <h2>.)

Example 5: 
http://www.dfat.gov.au/dept/annual_reports/02_03/overviews/2.html. Valid 
XHTML 1.0 Transitional 
Uses <b> instead of something semantic (e.g. <strong>),although the real 
problem with this page are missing longdescs for the charts and figures. 
Not the best example of non-semantic markup.

>- Missing metadata (<caption>, <tbody>)

Example 1: http://www.mathunion.org/mph/MPHMaker_help.html. Valid XTHML 1.0 
Transitional (see 
The table after the heading "2. The Content of Groups and Subgroups of the 
MPH" is a data table without <caption>, <thead>, <tbody> or <th>.

Example 2: 
Valid XHTML 1.0 Transitional (see 
This page contains a complex data table without <caption> and with all 
content (including cells that should be headings) in <tbody>.

Example 3: http://orgel.datzko.ch/examples.html. Valid XHTML 1.0 
Transitional. The page contains a data table with <tbody> and without 
<thead> or <th>. <caption> is also missing; the data table is nested inside 
a layout table with two rows; the first row contains one cell with the data 
table; the second row contains one cell that contains the "caption".

Example 4: http://www.lwl.ch/science/hermen/balustrade.html. Valid XHTML 
1.0 Strict. The page contains a data table; the "caption" is in a paragraph 
below the table (<p><strong>Tabelle 1:</strong> Masse der Hermen</p>) 
instead of in a <caption> element.

>- Missing functional code (<noframes>, frame titles)

Example 1: http://www.du.edu/uts/classes/advhtml/simpleframer.html. 
Tentatively validates as HTML 4.0 Frames (no character encoding found: 
There is a <noframes> (which points to an unframed version and contains a 
"Valid HTML 4.0!" logo), but frames have no title attribute (name 
attributes contain the values "tocFrame" and "docFrame").

Finding frame-based sites that actually validate is harder than I expected. 
Most of the frame-based sites that I found in the last few hours don't even 
have a DOCTYPE. Anyone else?

- Other issues (e.g. nested tables that validate):

Example of almost valid code but with nested layout tables: 
(validation error caused by "non SGML character number 1"; see 

>If you can't find three examples of each of these in a day's worth of 
>browsing, I'll eat my hat and both of my shoes.

I've almost saved someone from a severe case of indigestion :-)


Christophe Strobbe

>And that's just the stuff that's inside the angle brackets. Being valid 
>doesn't help any of it.

Christophe Strobbe
K.U.Leuven - Departement of Electrical Engineering - Research Group on 
Document Architectures
Kasteelpark Arenberg 10 - 3001 Leuven-Heverlee - BELGIUM
tel: +32 16 32 85 51
Received on Tuesday, 21 June 2005 14:51:43 UTC

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