RE: Moving Conformance Claim

No, P3P is a vocabulary for privacy. It would require writing a new
vocabulary - in essence repeating the work being done with EARL.

In order to make EARL a feasible approach, each checkpoint being conformned
to needs a URI.

This is not very difficult - and there are tools already doing this for WCAG
1.0 (AccessValet, Accverify, a bookmarklet from Jim Ley, among others).

It is possible to associate EARL information by several means - on the Web it
is possible to use annotations, or point to it from a page using the link
element, or to include it directly in some XML formats (not HTML unless you
write a customised schema for it, but in VoiceXML, SVG and SMIL it is
possible with the specifications as they are).

Hopefully by the end of this month I will be able to demonstrate a few simple
open-source tools for using EARL to do assessments. There is a project in
France which should make an open-source automated checking tool like Bobby
use EARL (this is what accessvalet already does, except that accessvalet
isn't free software - the maintenance vs adapatability trade-off is a topic
for somewhere else) in a few months.

I would seriously consider EARL, and getting some people who have real
expertise in it to talk to the working group before rejecting it or repeating
the work.

cheers

Chaals

On Mon, 14 Oct 2002, Lee Roberts wrote:

>Would it be possible to use the same format that
>P3P uses to accomplish this issue?
>
>Lee
>  -----Original Message-----
>[mailto:w3c-wai-gl-request@w3.org]On Behalf Of
>Gregg Vanderheiden
>  EARL was brought up as the logical approach.
>However, note that EARL is like XML in that it is
>a method for writing conformance statements.  It
>does not necessarily specify exactly how to
>write-up the conformance statement so that it
>would be uniform across sites and therefore
>searchable.
>
>
>  Also note that the statement may have to be on
>every single page in order for a search engine to
>be able to evaluate it based upon a “hit” on a
>page.  The inclusion, however, might be able to be
>indirect.

Received on Monday, 14 October 2002 12:37:47 UTC