Re: Validation as test for basic accessibility

[David Poehlman]
>bruce, I haven't an url, but another postulation.  when accessability is
>truly integrated into the spec, we won't have to worry about compliance,
>we can just validate.

I don't believe that that can ever happen.  Any document type includes
machine-checkable specifications and non-machine-checkable specifications.
Most accessibility guidelines fall into the latter category.  A machine can
check that all <img>s have alt attributes, for example, but it can't
(easily) check that the value of the attribute is useful for communicating
the author's intent behind the image.  Similarly, a computer can't check
that the linearization of a valid <table> produces a stream of text that is
sensible; these are value judgments best made by a human.

There will always be badly-written pages.  With sufficient educational
efforts, accessibility will be an automatic part of writing well, and only
the badly-written pages will be inacessible, but there will always be some.
Until true and complete AI comes along, we will have to accept that
accessibility depends on good intentions and kicks to the head.

-Chris

--
Christopher R. Maden, Solutions Architect
Exemplary Technologies
One Embarcadero Center, Ste. 2405
San Francisco, CA 94111

Received on Wednesday, 19 January 2000 13:53:53 UTC