WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: Yes!!! It's the Accessibility World Cup!

for

From: Jukka K. Korpela
Date: Jun 12, 2006 4:40AM


On Mon, 12 Jun 2006, Daniel Champion wrote:

>> validity in itself is not intrinsically a proof that accessibility is a
>> concern for any given site.
>
> I agree, that's why I specifically said it would be a reasonable *proxy*
> for accessibility.

That sounds like a confusing attempt to use the technical term "proxy"
metaphorically. Human communication usually fails, and metaphors make
failures more or less inevitable.

Validity is trivial syntactic correctness and it has very little to do
with accessibility. Some syntax errors imply serious problems in
accessibility, most don't, though they may affect other aspects of
quality. If a syntax error causes an accessibility problem, then that
particular error needs fixing.

> Developers of large sites who understand the value and
> benefits of valid HTML are also likely to understand and implement
> accessibility.

Hardly. This is just an assumption, and often proven wrong.

> Of course there will always be exceptions,

Exceptions to what? There are accessible sites that are valid. _That_ is
an exception.

> and of course
> valid code alone does not make a site accessible.

The point is that it has almost no impact on it. On a sunny day, someone
will somewhat improve accessibility by adding some alt attributes just
because validation made him realize the need for such attributes. Much
more often, he will insert silly alt attributes just to please a validator.

> The point I trying to make is that automated testing of accessibility is
> extremely limited without manual checking,

I couldn't agree more, except possibly if you changed "extremely limited"
to "useless" or "worse than useless", but that would perhaps be too
extremistics. Automated testing has a little use - in the hands of a very
competent author or tester - though mostly it just confuses people and
wastes their time.

> A validity
> league table would be no better at measuring accessibility, but it would
> be less ambiguous and misleading.

Data about validity would be objective but mostly irrelevant. I'm not sure
about the objective part either, due to problems in validators; the W3C
validator has known limitations in processing XML and XHTML.

> This email and any attachments have been scanned for viruses prior to
> leaving Clackmannanshire Council.

I guess that text got added independently of you, but have you reported
to the appropriate people that such statements are foolish and may confuse
some people and annoy others? (It annoys me because I see the gross
absurdity too well.)

--
Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/