W3C home > Mailing lists > Public > w3c-wai-au@w3.org > January to March 1999

HTML 3.2 vs 4.0 and def'n of accessibility - an issue

From: Charles McCathieNevile <charles@w3.org>
Date: Wed, 13 Jan 1999 15:03:02 -0500 (EST)
To: Rob Cumming <nugget@sausage.com.au>
cc: w3c-wai-au@w3.org
Message-ID: <Pine.LNX.4.04.9901131416220.4821-100000@tux.w3.org>
It is in a context like this that I think the AU Guidelines may need to
discuss accessibility explicitly, rather than bounce the problem to the
Page Authoring Guidelines.

For example, we can suggest to an individual author that they use HTML 4.0
and CSS2 wherever possible and leave it at that. But as Rob pointed out,
what Tool developers need is to know what 'best practise' represents, AND
know how to produce code which works for the people out there now.

As a rough start, I use the following algorithm:

HTML 4.0 (transitional) and CSS2 allow almost everything that is in HTLM
3.2. The exceptions are things like IMG without ALT which is allowed by
3.2 but not 4.0, and is also not helpful. (Note that ALT="" is not
necessarily wrong - there are contexts where it makes perfect sense for
accessibility)

Therefore, producing HTML 4.0 with CSS 2 will ensure that as newer
browsers are produced taking advantage of the features offered, both the
functionality and the appearance of your pages can be enhanced by
providing extra controls not currently available.

However for many of these, a 'legacy solution' is required. Often this is
a simple matter of code insertion. For example, to include a mathematical
expression there are a number of techniques.

One possibility for an Authoring tool which allowed the creation of
mathematical equations is to automatically generate output versions as
MathML, a GIF image of the equation, and a plain-text dscription. These
could then be included in an HTML 4.0/XML compliant way (except my XML
grammar is not too good) like:

<OBJECT TYPE="text/xml" DATA="equation1.mml" WIDTH="111" HEIGHT="44">
  <OBJECT TYPE="image/gif" DATA="equation1.gif" WIDTH="99" HEIGHT="44">
     <A HREF="equation1.html"><IMG SRC="equation1.gif" ALT="Equation
One"></A>
  </OBJECT>
</OBJECT>

The algorithm for producing this is fairly simple, and I imagine that in
most cases the hard work would be in encoding the multiple output systems.
Since I don't use mathematical software I don't know if they have done
that already, but remember this is a theoretical example to illustrate a
design principle. I realise that a text description of an image cannot be
generated automatically in the way that one can for a mathematical
equation. The gathering of descriptions is a seperate problem (although it
too needs to be addressed).

The output then works really nicely for people with a MathML compliant
browser, provides a GIF image of the equation for those using a standard
visual setup, and provides a simple link to a description of the object
for those using a standard non-visual setup.

It therefore provides a 'best-practise' solution, one which works for 'the
bulk of the market' and one which provides accessibility, without any
great hardship being undergone by the author of the page. (in my humble
opinion, of course)

I think that our techniques document, which it seems we have decided to
produce, will include many such example methods for integrating
best-practise, wide compatibility, and accessibility for 'the tiny
minority'. (Actually I think the tiny minority is much larger than
generally described. But statistics are just that, and generally
unavailable in any case. sigh.)

In particular I think that the use of CSS is amenable to such a treatment
(since I know this is done already in a number of large sites, and seems
to be the transition path most likely).

cheers

Charles


  At 05:08 PM 1/13/99 +1100, Rob Cumming wrote:
  >In this context HTML 3.2 still 
  >represents the level of markup that the
  >majority of our users want to work to, with perhaps a third of them 
  >seriously utilizing HTML 4.0 markup, perhaps a tenth
  >might dabble with CSS 1 and XML/CSS 2 dont even get on the radar.

And BK (I think) added
  
  Maybe we should add something allowing for Authoring Tools to have
  different levels of "checks." In the preferences, a user says they only
  want to program in HTML 3.2, (default being 4.0), and as they program, they
  are told when something is NOT accessible according to HTML 3.2...
  
  

--Charles McCathieNevile -  mailto:charles@w3.org
phone: * +1 (617) 258 0992 *  http://purl.oclc.org/net/charles
W3C Web Accessibility Initiative -  http://www.w3.org/WAI
545 Technology sq., Cambridge MA, USA
Received on Wednesday, 13 January 1999 15:03:15 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 22 September 2008 15:52:54 GMT