machine-readable claims Re: Conformance Ideas -- Collection #1

Hi Anne,

it isn't that the data isn't meant to get to the user. But just like HTML
code, it isn't presented in raw form, but interpreted by a machine.

There are several benefits of this, as outlined by others.

One is that it is then fairly easy to make the machine present the claim in
whatever language the user wants. As pictures, words, or both.

Another is that it is possible to search for exact claims - if we allow the
old proposed consensus statement C6, where it is possible to make partial
claims, even to the extent of being able to say "this page is illustrated
appropriately for a simple reading level, even though it lacks textual
equivalents (or vice versa)".

I hope this makes it clearer.

Technical digression: There are, and have been for years, search engines and
browsers (such as IE and Netscape) that can use PICS ratings to filter
content. There was a basic PICS scheme developed for assessing conformance to
the guidelines, but it was not promoted well within the group let alone
outside, so there is litlle content that uses it. There has now been a lot of
work done on a system called RDF - Resource Description Framework. This
allows description of resources (which really means things with URIs, such as
Web pages or parts of them) in a way that means the computer can interpret it
and present it to people afterwards, and more usefully, can add a few bits of
information together and make basic conclusions first (like "this page has
text equivalents, but they all say 'image or media object' so it isn't likely
to be useful to a text-based user"). Using this information in the popular
search engines is in its infancy, but is a major project of the US
government, the European Union, W3C, and others.

The Authoring Tool group, the Evaluation and Repair tool group, and others,
have been working on this kind of evaluation already, and are now at the
stage of deploying it  in tools - there are demonstrations available, and we
are hopeful that you will be able to see it in major market tools within the
next year (although you won't "see" it if it works, the tools will just be
smarter about helping you).

It would be possible to use PICS and therefore existing systems to provide
the kind of detailed claims talked about above, but it is somewhat harder
than with RDF both to make the statement and to process it. Given that in
WCAG 1.0 the take-up of machine-readable claims was fairly poor it didn't
seem worth the effort of pushing to put in place a system that was already
obsolete.

cheers

chaals

On Tue, 9 Oct 2001, Anne Pemberton wrote:

  Jim,

           So, if not the user, then who is the audience of the "conformance
  data"?  the regulating agencies?

                                                   Anne

Received on Wednesday, 10 October 2001 03:46:02 UTC