Re: persistent QA problems with the W3C Validator

Charles McCathieNevile <> wrote:

>I'd just like to second Gregory's comments about soliciting feedback on
> and the Interest Group list.

Thank you (both) for the pointer. I will surely take advantage of it.

The reason why I haven't done this earlier is simply that the code is in no
fit state to make these changes. The HTML is so intermingled with the code
that any trivial change can turn out to be a monumental task. Asking the
experts to review it only to be unable to do anything about it strikes me
as singularily rude and a waste of everyone's time. I have made changes
here as circumstances have allowed -- mainly due to the greatly appreciated
efforts of the people I mentioned in my previous message -- but a lot has
been put off for the reasons above.

This is one of the reasons why I've put my efforts into a major ovehaul of
the Validator code over the last year or so. My goal is to move all the
HTML out into separate (template) files so that you don't need to
understand 2000+ lines of grotty Perl code to be able to make Accessibility

That work has begun to pay off. Some such changes are now trivial to make
(allthough much yet remains), and I was recently able to add XML, EARL/RDF,
and N3 output options to the Validator with little effort (many many thanks
again to Sean Palmer, Jim Ley, and Nick Kew for their invaluable help with
this!). This suggest to me that I'm on the right track; concentrate on
fixing the infrastructure instead of spending time on workarounds and
bandaids, because once the infrastructure is properly fixed the rest will
follow naturally.

>PS As far as I can tell W3C doesn't have any coffers to support thias
>kind of thing. Sigh. I am sure they accept donations.

Hmmm, perhaps this would be an issue for the QA Activity. Making sure the
W3C's own pages are Accessible seems a natural element of Quality
Assurance. The Recommendations _must_ be accessible, but it would be
natural to also have the various WG, Activity, and Area home pages hold to
some such standard.  QArl?

Received on Sunday, 9 June 2002 10:35:32 UTC