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

16 Jan 2001 UAAG 1.0 Guidelines and Techniques availability

From: Ian Jacobs <ij@w3.org>
Date: Tue, 16 Jan 2001 23:42:58 -0500
Message-ID: <3A6522D2.5B04722A@w3.org>
To: w3c-wai-ua@w3.org
Hello,

Surprise! I've published the 16 January 2001 version of the
Guidelines [1] (the techniques [2] are unchanged). 

I published this version because I was unhappy with the
conformance model of the 13 January 2001 draft. 

There are a few editorial changes in the document based
on some comments from Eric Hansen, but the primary change
is that the conformance model has been revised. While it
resembles the previous versions, I think it gets rid of
some bugs. I encourage review of section 3 (Conformance [3])
because I want to be sure that the Working Group agrees with
this approach.

Here is the problem: WCAG 1.0 (by comparison) has only
three ways to conform: A, Double-A, or Triple-A. We have
a lot more permutations because we allow variable support
for content types and input modalities, in addition to 
the A/Double-A/Triple-A conformance levels. You count the
permutations:

   3 Conformance levels times
   7 Content type labels times
   2 Input modality labels

I believe that's 42 permutations, and you can add 
"applicability" to that...

We don't have 42+ different labels to indicate conformance,
so this version of the document introduces only two:
unconditional conformance (the user agent 
satisfies all the requirements) and conditional conformance 
(the user agent satisfies fewer than all requirements, in
a well-defined manner). However, with that means that 
two user agents that conform conditionally may differ greatly
in the requirements they satisfy. 

The only reason I'm willing to live with that situation is
that any valid conformance claim MUST indicate which requirements
the user agent does not meet. It does this through 
conformance levels, content type labels, input modality labels, 
and assertions about applicability. 

Thus, two user agents may conform conditionally and satisfy
very different requirements. But two claims of conformance
must state these differences in detail.

Last big change: the definition of a "valid claim" now includes 
three conditions:

  1.The claim is well-formed. 
  2.The claim indicates which requirements the user 
    agent does not satisfy through one conformance level and any
    relevant content type labels, input modality labels, and 
    applicability information. 
  3.It is verified that the user agent satisfies all other 
    requirements not exempted by the claim through these
    mechanisms. 

The third condition introduces the word "verified" for the first
time. The document doesn't explain HOW one verifies that a 
requirement has been satisfied, but that's always been a limitation
of the document. (Recall that WAI doesn't certify claims today either.)

Comments welcome,

 - Ian

P.S. I don't intend to publish a new draft every three days.
If the conformance model is still broken, I will not republish
immediately to correct it.

[1] http://www.w3.org/WAI/UA/WD-UAAG10-20010116/
[2] http://www.w3.org/WAI/UA/WD-UAAG10-TECHS-20010116/
[3] http://www.w3.org/WAI/UA/WD-UAAG10-20010116/#Conformance
-- 
Ian Jacobs (jacobs@w3.org)   http://www.w3.org/People/Jacobs
Tel:                         +1 831 457-2842
Cell:                        +1 917 450-8783
Received on Tuesday, 16 January 2001 23:43:06 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:49:29 UTC