w3.org site-wide markup review?

For reasons I forget, I'm subscribed to site-comments@w3.org, archived
at [1]. Aside from a recent flurry about how wonderful the new HTML5
material is (and requests for stickers - me too!) the majority of
messages seem to be about questionable markup on the site. I don't
know what processes are already in place for checking the
accessibility and usability of pages, but that there are any messages
of this nature suggests that things aren't quite as joined-up as they
should be in W3C-land.

Ok, there are things that will slip through any net. The use of CSS
fixed px font sizes seems to be an example, it doesn't seem to be
checked by the online validators I tried (applied to the page
http://www.w3.org/Amaya/ - though contrast issues were flagged). But
given the W3C's key role in producing the relevant specs and
guidelines, there's a good case for saying its own pages should be
subject to far higher standards of quality control than any other on
the Web. Best practices, leading by example and all that.

A good way for dealing with this would be for the W3C to instigate an
independent review, and to put automated processes* in place to ensure
continuing quality of material. Ok, such things would cost non-trivial
time & money, but even if the point of principle wasn't enough, the
surprising amount of hostility in some of the messages to
site-comments extrapolates to much wider, unvoiced, annoyance or at
least dissatisfaction. i.e. this is a credibility issue, very bad for
PR.

Whatever, perhaps there's a cheaper solution. I'm guessing there are
plenty of companies working in the WAI space with products to sell. If
one were to apply their tooling to the w3.org site, it would be a
great demonstration for them - and maybe they could be given some
appropriate stickers :)

Any takers?

Cheers,
Danny.

* automated process - not rocket science, I bet the necessary kit is
around nearby, might even already be assembled (but no doubt in need
of updating). I reckon it would need the following:

1. (a quick review of the EARL vocab)
2. a triplestore (an online one with SPARQL endpoint would be good transparency)
3. a dataset listing individuals/groups responsible for the various
areas of the W3C site (and maintainers of tools like spec-doc
generators)
4. a HTML, CSS, RDF (and any other relevant formats) validator and a
fine-grained, ultra-sensitive checker (some kind of fussy lint)
5. a spider hooked up to 4. pumping EARL data into 3.
6. a bugtrack/notification system, sending reports to the people in 3.
*and* confirming action is taken


[1] http://lists.w3.org/Archives/Public/site-comments/

-- 
http://danny.ayers.name

Received on Tuesday, 1 February 2011 09:50:05 UTC