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

RE: Problems with guideline 4.1

From: Jason White <jasonw@ariel.ucs.unimelb.edu.au>
Date: Mon, 26 Jul 2004 11:03:48 +1000
Message-ID: <16644.22644.735504.363192@jdc.local>
To: Web Content Guidelines <w3c-wai-gl@w3.org>

Gregg Vanderheiden writes:
 > I think I disagree with 3.  if you have a URI that no one but you can access
 > then no claim will be made against it - or it doesn't matter.  If it is
 > rendered before viewing then it passes.  It is what is delivered that is
 > tested.    

I agree with everything up to the last sentence, and that's where
there might be a problem. If we apply the "one conformant version must
be available to user agents" rule (see the "authored unit" thread)
this could solve the problem. Then we say that conformant versions
must be delivered to user agents in a format that conforms to

Suppose however that I leave it up to third parties to do the
conversion from my XML format into formats suitable for user agents -
i.e., I am supplying a Web service. Does this mean I can't make a
conformance claim with respect to my content because I don't meet
guideline 4.2? If so, then by the aggregation rule in our latest
conformance proposal the third parties who transform and deliver my
content can't make conformance claims either, because the conformance
of the whole is equal to the lowest conformance level of any of the
parts. If the third parties are transforming rather than aggregating
this may not apply - but what's the difference between the two?
 > RE 5 - I disagree.  Conforming to spec is a good thing - but this is only
 > about access.  It can't be level 2 unless you can show that conformance has
 > a big increase in accessibiity -- AFTER all of the other guidelines have
 > already been met.

I think it's a problem, not so much for end users but for the
designers of tools. For example, consider a set of XSLT style sheets
designed to process (X)HTML in various ways to improve its
accessibility. This is not just a hypothetical example; such style
sheets are supplied with Emacspeak. Now if the content is invalid,
i.e., doesn't conform to specification, then one can't simply apply
these style sheets to get the desired results, or at the very least
their design becomes considerably more complicated. Ultimately one has
to resort to complicated heuristics to discern the structure of the
content and process it accordingly. So, yes, I would say that
invalidity places a heavy burden on tool/style sheet designers and
hence limits their availability as means of transforming content to
suit the access needs of users.
Received on Sunday, 25 July 2004 21:03:55 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 16 January 2018 15:33:50 UTC