WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: HTML Validator and CythiaSays


From: Jukka K. Korpela
Date: Oct 30, 2003 11:25PM

On Fri, 31 Oct 2003, John Britsios wrote:

> I also wanted to say that there is more than one definition of the word
> validator.

Surely. But in the context of SGML applications such as HTML, only one.

> CSE HTML Validator has never sold itself as a "formal" DTD validator.

That sounds like an ex-president not having had sex "formally".

> It is a validator in the common definition of the word, the definition
> that most people use.

What definition? Most people have no definition for that word. HTML
authors should know the definition that is relevant in the HTML context,
and if you are selling something as HTML validator that isn't a
validator, you are misleading them - in this context, for commercial
purposes. I understand well that the name "CSE HTML Validator" has
commercial value as a trademark. This does not make it less misleading.
(Besides, it's both less and more than a validator; a large part of
what it does doesn't really deal with markup syntax at all.)

"CSE HTML Validator" has many interesting features, and it could even be
useful in the hands of a very competent author who knows well the
limitations and errors of the software and uses it as tool with a very
critical mind.

But it is isn't useful to a novice, or to an author with little
experience, and it isn't a validator _even_ a very loose sense of the
word. Regarding accessibility, there can be no such thing as validator
even in the loose sense, a program that reports whether or not a document
satisfies some accessibility criteria (which is a meaning _quite_
different from a markup validator). An _invalidator_ is possible, i.e.
a program that is able to correctly report for some documents that a
document does not comply with some accessibility criteria,
but even that is difficult since the criteria themselves are so vague.

And in accessibility, "CSE HTML Validator" doesn't seem to do anything
that the free tools wouldn't do, except give some _wrong_ diagnostics.
Most accessibility checkers give wrong diagnostics among useful messages,
partly because the accessibility criteria weren't really designed for
automatic checkability (which is a good thing, since if they had been,
they would have a very narrow scope and limited impact on actual

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

To subscribe, unsubscribe, suspend, or view list archives,
visit http://www.webaim.org/discussion/