Comments on guideline 13

In relation to guideline 13:

1. Despite insistance by Charles McCathie-Nevile and myself that authors
comply with document type definitions as provided in W3C specifications,
this requirement seems to have disappeared from the guidelines as some
stage during the long revision history of the document. While guideline 13
refers to using W3C technologies according to specification, the
requirement for valid DTD compliance, which used to be given in a
checkpoint as I recall, is no longer there; and nor is it mentioned in the
Techniques document. This omission ought to be rectified.

2. Checkpoint 13.2? mandates avoidance of deprecated language features. I
would suggest removing the qualifying words "whenever possible" from this
checkpoint, as it is in principle always possible to avoid such deprecated
features. I would also recommend considering in which other contexts
qualifications such s "whenever possible", which tend to eroide the
importance of the checkpoint, can be removed; the priority is reflected in
the rating of the checkpoint (P1, P2 or P3) and it is unnecessary to add
unnecessary confusion by further limiting the requirement with such
ambiguous qualifications. More specifically, the expression "whenever
possible" implies that the author is to make a judgment as to whether a
particular feature is widely enough supported, or whether the results of
applying the suggested technique would have sufficiently significant
backward compatibility problems to warrant non-compliance with the
checkpoint. These considerations should be made explicit, as has been done
elsewhere in the guidelines with regard to tables, frames etc., instead of
using a generic phrase such as "whenever possible", which implies that an
author can take an indeterminate range of considerations into account in
choosing whether or not to apply a technique, this being altogether
contrary to the aim of the priority ratings, which are to be judged
entirely on the basis of impact (the degree to which an accessibility
barrier results) rather than on convenience for the author, backward
compatibility, expense, or other such factors. The same comments apply to
checkpoint 13.1 and probably also to other uses of the "whenever possible"
exemption.

Received on Monday, 1 March 1999 18:32:47 UTC