Organization of WCAG 2.0 and checklists

In commenting on WCAG 2.0, Terry Thompson suggested that the W3C
Recommendation should contain only the normative parts of the
guidelines document, and that the non-normative  (explanations,
examples etc.) should reside in a separate document, connected to the
guidelines via links.

I disagree with this proposal, mainly because it is completely out of
line with W3C practice, including the design of other guidelines
documents. I think there ought to be some uniformity in structure,
organization and, where possible, terminology across W3C guidelines as
such. Also, in order to understand the guidelines it is often
desirable, even if not strictly necessary, to read the explanatory
notes and examples; separating them from the principal document would
greatly reduce their visibility to first-time readers.

As an alternative I suggest:

1. that as in WCAG 1.0 we create a checklist, in this case comprising
   only guidelines, checkpoints and success criteria, to be published
   with the guidelines as part of the W3C Recommendation.

2. that a link to this checklist be included in the principal
   guidelines document, as close to the beginning as is possible while
   conforming to W3C publication policy.

3. As Gregg? suggested, it might also be useful to provide such a
   checklist in the form of a table. If so, this should be included as
   well.

I am hopeful that a solution of this kind (with or without
modifications) will be judged satisfactory.
It also eliminates the problem of having four types of document:
guidelines, guidelines +comments/examples, techniques and checklists.
On the present proposal there would only be guidelines, techniques and
checklists (including one checklist linked near the start of the
guidelines themselves).

Further, I like of Andi's suggestion of a web page that gives links to all
of the checklists and techniques documents as well as other relevant
publications. Actually, the checklist page may amount to a form
through which the dynamic generation process can be controlled, but
that is a decision yet to be taken.

Received on Thursday, 31 October 2002 03:22:50 UTC