W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > January to March 2005

RE: Guideline 4.2 and UAAG

From: Jason White <jasonw@ariel.its.unimelb.edu.au>
Date: Mon, 14 Mar 2005 19:46:50 +1100
Message-ID: <16949.20346.967588.95902@jdc.local>
To: "John M Slatin" <john_slatin@austin.utexas.edu>
Cc: "Loretta Guarino Reid" <lguarino@adobe.com>, <w3c-wai-gl@w3.org>

John M Slatin writes:
 > David voiced a similar concern the other day, wondering whether adopting
 > UAAG as baseline would amount to wrapping the entire WCAG 2.0 inside an
 > Until user agents clause... (<untilUserAgents>WCAG
 > 2.0</untilUserAgents>).  Jason's response was compelling:
 > Jason argued that this would not be the case. In Jason's view, adopting
 > UAAG as baseline would address the underlying issue that Roberto
 > raised: it would allow (or force) us to make a cleaner separation
 > between what belongs to guidelines for the accessibility of Web
 > *content* (the purview of WCAG) and what belongs to user agents (the
 > purview of UAAG).

Precisely. Thank you for a clear and cogent summary of the dialogue, John.
 > The theory makes a great deal of sense to me. It would mean, for
 > example, providing some kind of "fallback" or "repair" technique.
 > Because the techniques documents are non-normative, there *could* be
 > "Until user agents..." techniques.
Exactly, but the proposal as I understand it is that there would be no
"until user agents" success criteria.
 > It's not clear to me, however, whether we would be saying that use of a
 > fallback or repair technique would constitute conformance in the absence
 > of a conforming user agent. If the answer is "yes," you would be allowed
 > to claim conformance if there's no alternative to using a
 > fallback/repair technique, I'm not sure I see how we've escaped the
 > "Until user agents..." trap in anything but theory.  And if the answer
 > is "No," using fallback techniques means you can't claim conformance,
 > then it seems to me we *will* have wrapped the Guidelines in an "Until
 > user agents..." clause.

One solution here was hinted at in Gregg's suggestion of adding, where
relevant, wording to the success criteria requiring that the content
be implemented "in a standard and supported manner". A definition
could then be given of what "standard" and "supported" require, which
would be stated with sufficient generality to create an evolving
standard. This could even be stratified into three levels, as in the
guidelines themselves:

Level 1: there is an implementation that supports this technique and
it meets some basic standard of interaction with various kinds of user
interface or assistive technologies.

Level 2: More user agent implementations are available that meet UAAG
(or whatever we decide the requirement should be).

Level 3: This is really ubiquitous and well supported.

The division into three levels is of course optional, but it would
satisfy those who think there need only be a fairly minimal
implementation as a prerequisite to an author's relying on a technique
by offering minimal (level 1) conformance under those circumstances,
while allowing higher conformance levels to be claimed to distinguish
content that is truly backward-compatible.
Received on Monday, 14 March 2005 08:48:33 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 16 January 2018 15:33:52 UTC