WCAG2 draft - question and (minor) comments

Dear WCAG Working Group.

I have had a chance to look at your latest (public) draft. It is really 
a very good document, very clear and pleasant to read, my sincere 
congratulations for your work so far.

The QA working group ( http://www.w3.org/QA/WG ) is currently authoring 
a family of guidelines documents ( http://www.w3.org/QA/WG/#docs ) and 
will certainly be interested in your answers to the following question:

It seems that your group has recently chosen to change conformance 
criteria scheme from using priorities to using levels. It is an 
interesting move that the QAWG would like to analyse (both with regards 
to the guidelines being written by the group and with regards to 
specification authoring in general - since this is the scope of one of 
the documents in the QA framework).

I think that naming level 1 "MINIMUM" is a good move towards clarity 
(which is probably a partial answer to the question), and the 
"conformance" section of the requirements document for WCAG2 also 
partly answers the question, but the QAWG would be interested in 
knowing the full rationale behind this choice.


*******

And now a few comments about the draft... (unfortunately partly written 
offline, with only a printout of the spec, so please forgive me for any 
mistake that I may make).
These comments are mine, but some of them come from an analysis of your 
draft against the QAWG's "Checklist for Specification Guidelines" ( 
http://www.w3.org/TR/qaframe-spec/qaframe-spec-checklist.html ).

- A first question is about the (lack of) use of RFC 2119 vocabulary to 
tag strict and optional conformance to your guidelines. I believe you 
did a good job of tagging strict and optional requirements without the 
use of MUST MAY etc, and this mechanism seems to fullfill your 
requirement to be testable, although there is (correct me if I'm wrong) 
no machine-readable version of your guidelines and the spec is not 
authored using a structured grammar such as XMLSpec.

In the absence of those (RFC 2119 vocab, machine-readable markup) I 
think your conformance section, albeit well written, would benefit from 
an explanation of this choice (maybe with a mapping of your 
presentation vs a "usual" MUST/MAY example).

- Speaking of the conformance section, I have a small suggestion :
for the sake of clarity,
"you will have successfully met Cp.X at the minimum level if...
you will have successfully met Cp.X at level 2 if
	- foo
"
could read
"you will have successfully met Cp.X at the minimum level if...
you will have successfully met Cp.X at level 2 if
	- you have met Cp.X at the minimum level, and
	- foo
"
I know that the conformance section is clear on this point, but I would 
however consider the proposed verbiage to be even clearer about the 
relation between levels.


- In checkpoint 3.2, level 2, the use of "display" for any rendering 
device (visual, audio, ...) seemed to me as as potentially misleading. 
I may have missed other uses of "display" as a generic rendering 
device, or a definition of "display", but I think it would be better to 
not use "display" for any other context than visual.

- The title for checkpoint 3.6 talks about "graceful recovery" from 
errors but the text for the 3 levels do not seem to include this 
concept. It may be a good idea to introduce the idea of "fallback 
behaviour", etc.

- many guidelines are still quite HTML oriented, e.g "navigation bars" 
in Cp3.4. If one of the goals of WCAG2.0 is to leave HTML for more 
generic grounds, it might be necessary to move html specific checkpoint 
wording to the examples.


Thank you.
olivier.
-- 
Olivier Thereaux - W3C - QA : http://ww.w3.org/QA/
http://www.w3.org/People/olivier | http://yoda.zoy.org

Received on Friday, 18 October 2002 08:10:03 UTC