W3C home > Mailing lists > Public > w3c-wai-ua@w3.org > April to June 2001

Re: UAAG Conformance Example In EARL

From: Sean B. Palmer <sean@mysterylights.com>
Date: Tue, 15 May 2001 01:16:28 +0100
Message-ID: <067401c0dcd4$959f5ac0$58d893c3@z5n9x1>
To: "Ian Jacobs" <ij@w3.org>
Cc: <w3c-wai-ua@w3.org>, <w3c-wai-er-ig@w3.org>
Hi Ian,

> We are still working on the requirements for well-formed
> claims, in particular to allow for increased granularity.

Scope, obviously, seems to be the most important factor here - in as
much as UA are obviously expecting these conformance claims to *do*
something. I presume that this is (primary) to provide information to
the user about how accessible a product is, and (secondary) to assist
in development and debugging of the product itself. As far as
"well-formed" is concerned, the modality of the information presented
is perhaps not as essential as the content itself, but it is certainly
a huge factor for consideration. XHTML "as is" is great for
documentation, but next to useless when it comes to machine
processing. EARL, on the other hand, isn't 100% human readable, but is
an exemplary format for machine processing. Because it's cast in RDF,
it's also repurposable, and hence it may actually be possible to
transform a well-formed conformance claim in EARL into an XHTML
version.

If I beg to the ERT staff (my XSLT skills probably aren't up to it),
we might be able to get an example set up - transforming some EARL to
XHTML on the fly using a Makefile. EARL, because of it's scope (which
is to make extremely well grained evaluations), would probably be
quite useful as far as UAAG conformance is concerned.

> I like the use of rdfs:comment. That will also be important
> for checkpoints that the claimant says do not apply (for
> rationale).

Actually, that was a bit of a hack on my part: I should have looked up
the particular ID for the checkpoints myself, and pointed to them
using earl:id or something (we need to clarify the use of terms in the
next version of the schema). The addition of the earl:excludes term is
certainly quite useful for omitting checkpoints form a suite that one
does not conform to.

> [...] We require that one version of a claim conforms to
> WCAG 1.0 [...] A machine-readable claim such as one
> in N3 probably does not conform to WCAG 1.0, but is
> very useful (and probably sufficient). Do you have any
> comment on the accessibility of N3 claims?

I would claim, because of the repurposability of RDF, that N3 is
easily WCAG 1.0 compliant; perhaps more so than vague prose
descriptions, because we can be very explicit about which part means
what. You could only do that using XHTML if you had a bunch of
classes, and even then, because of the nature of the class mechanism
in XHTML, this would not be possible to describe to a high semantic
level. In other words, XHTML is really for prose, not data.

ERT is a WAI group, and we are looking more closely at using EARL for
accessibility evaluations than any other domain of technology at this
time. I think if there were any accessibility barriers in EARL, this
would have been pointed out quite some time ago. As ever, the defining
point is where one converts from data => prose; you don't want to lose
information, and you want to keep it accessible.

> In our conformance section, should we start pointing
> to EARL as a format for making claims?

I think you might want to start referring to it as a possibility for
making claims, yes; as informative information for the time being.
Once we have a stable base of implementation and so forth, we may
start asking if groups could provide normative links to EARL as a
method for expressing machine-readable evaluations. I'm sure AU and GL
will be looking into this as well (indeed, I know they are!).

--
Kindest Regards,
Sean B. Palmer
@prefix : <http://webns.net/roughterms/> .
:Sean :hasHomepage <http://purl.org/net/sbp/> .
Received on Monday, 14 May 2001 20:17:34 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 27 October 2009 06:50:50 GMT