RE: Proposal: Simplified presentation of conformance model

Sounds reasonable to me. See below for a comment (labeled "EH:").

-----Original Message-----
From: Ian Jacobs
To: w3c-wai-ua@w3.org
Sent: 12/28/00 8:38 AM
Subject: Proposal: Simplified presentation of conformance model

Hello,

Don't be scared off by the word "conformance" in the
subject of this message. I believe this proposal simplifies
the UAAG 1.0 conformance scheme. It does not stray far from
what we have already defined, and is, I believe, a clearer
model.

The model works as follows:

1) By default, a user agent must satisfy all checkpoints
   in the document in order to conform.

2) It is possible to reduce the "scope of a claim" [new term]
   in three ways:

   a) Specifying a conformance level. A conformance level is
      a labeled subset of checkpoints.
   b) Specifying one or more content type labels (more labeled
      subsets). 
   c) Specifying input modality labels (still more labeled
      subsets). [Note: Input modality labels still a work in
      progress...]

   (These are the three mechanisms we have chosen to
    increase the granularity of conformance claims.)

3) By default, all checkpoints *within the scope of a claim*
   apply. A checkpoint that is in scope does not apply 
   if any of three conditions are met. (Note: We may delete one
   based on some last call comments.)

EH:

I presume that the following are the three conditions:

From 23 Oct 2000 draft:

"The checkpoint makes requirements for graphical user interfaces or
graphical viewports and the subject only has an audio or tactile user
interface or viewports."

"The checkpoint refers to a role of content (e.g., transcript, caption, text
equivalent, etc.) that the subject does not recognize. For instance, HTML
user agents can recognize "alt", OBJECT content, or NOFRAMES content as
providing equivalents for other content since these are specified by the
markup language. HTML user agents are not expected to recognize that a text
description embedded without indicative markup in a nearby paragraph is a
text equivalent for the image."

"The checkpoint requires control of content properties (e.g., video or
animation rate) that the subject cannot control (e.g., the format does not
allow it) or does not recognize (e.g., because the property is controlled by
a script in a manner that the subject cannot recognize)."
==
IJ:

Why distinguish a checkpoint that is out of scope 
from one that does not apply?

 - A checkpoint that is out of scope is simply not part of a 
 claim. The claim makes *no statement* about whether this 
 checkpoint is satisfied or not. It is possible, for example,
 that a user agent would satisfy some, but not all, checkpoints
 for the "video" content type label. The label "video" would not
 appear in the claim, but the user agent would still satisfy 
 some of the checkpoints.

 - A checkpoint that doesn't apply *is* part of a claim. The 
   claim states explicitly that the user agent *does not* 
   satisfy this checkpoint (for if it did, then the checkpoint 
   would clearly apply).

I think it's worthwhile making this distinction. I also think 
that doing so simplifies the presentation of the material and
the definition of applicable.

Continuing the model (essentially unchanged):

4) A well-formed claim must include information about both
   the scope of the claim, and for those checkpoints in scope,
   which ones are not considered to apply. (A well-formed claim
   must also include information about the version of UAAG
   and the date of the claim.)

5) A claim is valid if:
   a) Well-formed, and
   b) The subject of the claim as a whole satisfies all 
      the applicable checkpoints that are in the scope 
      of the claim.

That's the model. Because I do not consider this a substantial
change from the existing conformance model (only a clarification
and simplification), I intend to integrate this into the next draft 
for Working Group review. 

 - Ian
 
-- 
Ian Jacobs (jacobs@w3.org)   http://www.w3.org/People/Jacobs
Tel:                         +1 831 457-2842
Cell:                        +1 917 450-8783

Received on Thursday, 28 December 2000 13:01:20 UTC