Re: Conformance Ideas -- Collection #1

From: Charles McCathieNevile <charles@w3.org>
Date: Mon, 8 Oct 2001 08:17:26 -0400 (EDT)
To: Gregg Vanderheiden <gv@trace.wisc.edu>
cc: "GLWAI Guidelines WG (GL - WAI Guidelines WG)" <w3c-wai-gl@w3.org>
Message-ID: <Pine.LNX.4.30.0110080806160.10980-100000@tux.w3.org>
```My ideas are scattered throughout. I am not treating these as three proposals
but as collections of individual ideas.

chaals

1.1.  We have layers of conformance (A, Double A, Triple A).

CMN Agree, if these are based on some rough idea of user needs

1.2.  You cannot claim level of conformance below A.  It is the
minimum.

CMN Agree.

1.3.  It is not clear whether only normative items would be used to
determine Double A or Triple A.

CMN There should be levels of conformance based on covering normative
requirements. It might make sense to have additions for including
non-normative requirements, but we don't have any of those yet so we should
cross that bridge later.

1.4.  Individuals doing more than A would claim A+.  Clicking on the +
would take them to a list of the items covered by the +.  1.5.  A
possibility would be to claim A+7, where a number would follow the plus
sign.  This would provide additional incentive for people to do more than
one or a few beyond A (since even one item beyond A would be A +, so why
do any more until you could get to Double A otherwise).

CMN This is a nice idea. But I think it is important to have fairly simple
levels of conformance - if there are too many it becomes a bit unliekly that
people will understand them, so they won't use them. (With A = barely, double
A = acceptable, triple A = useful we already have that problem...)

1.6.  All items are currently self-report, but normative items are
testable.

CMN I believe that we have a consensus requirement (again - it was also in
the initial requirements for WCAG 2) that normative items are testable. Which
means that this is really about whether or not conformance claims are
self-reporting. In the absence of a service provider to do this I think it is
the only possible approach.

1.7.  Conformance to informative items would simply be by assertion.
People conform if they assert that they conform.  Items should be worded
such that this makes the most sense.  This sounds problematic.  [Is
there an alternative?]

CMN I do not think it should be possible to claim conformance to some
non-normative requirement.

2.1. The working group should define one or more "standard formats" to
be followed in making conformance assertions.

CMN Agree

2.2. More than one format may be necessary due to the diverse
technologies which may be used to construct web content. For
example, text accompanied by a raster image icon may be the norm in
HTML, but inappropriate in SVG, PDF or other formats.

CMN This isn't a proposal.

2.3. The working group should permit EARL (the Evaluation and Report
Language) to be used to make conformance claims, in addition to or
in place of a human-readable conformance assertion.

CMN Agree.

3.1    A conformance scheme should be a meta-conformance scheme that
allows policy makers to describe their policy in WCAG 2.0 (and gives
guidance to minimum policy requirements), instead of serving as a policy
itself.  WCAG 2.0 conformance should not look like policy, but should
look like a toolkit for building policy.  This philosophy is in line
with the goals and aims of WCAG working group charter, as we are not
"writing laws" but we are writing primary material to be used by policy
setters (as well as providing technical documentation for developers).

CMN A conformance scheme should state clearly that it is conformance to a
technical specification, and that it should be understood as such. It might
be valuable to give examples of how different organisations would build
implementation strategies, and therefore intermediate local policies,
differently, although I would argue that this is work that belongs to EO and
we should ask them to accept it as a work item.

chaals
```
Received on Monday, 8 October 2001 08:17:27 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:47:16 GMT