E-mail List Archives

Re: Well formed verses Valid code

for

From: Andrew Kirkpatrick
Date: Feb 26, 2007 9:10PM


John,

> information highway, we're supplying the fuel. You can drive
> any vehicle you want, but the fuel I dispense will always be
> "pure" (valid).

Car engines, like browsers, are forgiving of impurities to a certain
extent. You can argue that cars would get better mileage if all fuel
was 100% pure, but the reality in the way that fuel is produced that
this is not always practical.

> I've
> seen code that rendered fine on screen, that read fine to
> screen readers, but was functionally "broken" simply due to
> invalid code (I'm thinking of some .net crap that only works

Maybe I'm missing something - it worked for AT and looked fine, but was
broken functionally?

> 2. Accessibility should be part of "The Policy" (full stop).

Of course.

> I am troubled that
> we again want to marginalize "accessibility" and speak about
> specifics regarding Adaptive Technology - yet at the same
> time we preach that Universal Accessibility is foundational,
> and not something that you bolt on at the end of the day.

I'm not sure which 'we' you're talking about. I'm not advocating any
interest in marginalizing accessibility. All that I'm stating is that
validity does not equal accessibility. You can take a perfectly valid
and accessible page and make a minor change (un-encode an ampersand in a
parameter string or use <b> instead of <strong> in an XHTML strict
document) -- is the page going to be less accessible? Keep making
changes and you will eventually affect the visual appearance and also
the experience for keyboard or screen reader or other AT users, but
those are the issues that we need to address and have easy to follow
guidelines for. There is just too much extra stuff that validity covers
that doesn't affect accessibility for actual users.

My responses are largely in response to Alastair's "I'm convinced that
well-formed (X)HTML should be an accessibility (as well as good
practice) requirement" statement. I'm not convinced that it should be
an accessibility requirement, but I am convinced that a robust
development process will address validity issues for other reasons.

AWK