W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > October to December 2003

RE: New proposal for WCAG conformance

From: Joe Clark <joeclark@joeclark.org>
Date: Thu, 6 Nov 2003 20:40:24 -0500 (EST)
To: WAI-GL <w3c-wai-gl@w3.org>
Message-ID: <Pine.BSO.4.53.0310181429130.27387@mail.veldt.ca>

> It addresses the following concerns
>
> 1 - multiple dimensions of conformance

Why do we need more than two *levels* of conformance?

> 2 - match to WCAG 1.0 in structure and measures

WCAG 1.0 is significantly broken in "structure and measures." If we match
2.0 to 1.0 too closely, then 2.0 ends up significantly broken, too.

> [BBC] I'd reword #2 to say "compatibility with WCAG 1.0 and other WAI
> guidelines" 3 - ability to both have minimum and advanced ways to address
> individual guidelines

i.e., we need only two levels, Core and Extended.

> 1 - that the 4 guidelines are now principles (Perceivable, Operable,
> Understandable and Robust)

though WAI, as ever, uses the longest, hardest-to-understand variants of
those concepts. Why can't they be nouns?

> 3 - that the success criteria remain success criteria  (the term checkpoints
> would be reserved for the checklists so we don't have confusion between the
> two which is almost certain if we have checklists and checkpoints that are
> different.

Yes. And so we don't end up with checklists of checkpoints, a term
understandable only to WAI insiders.

> - The terms CORE and EXTENDED are dropped. (these terms are used differently
> in other technical specs and didn't quite fit here.  Also they introduced a
> completely different structure and terminology than WCAG 1.0)

which is a *benefit*. The two-layer approach is massively simpler to
understand than 1.0, and also easier to *communicate*.

>     Category 1 = success criteria that
>
>        a) the author is not told how to present their information,

Adding an alt text *is* telling the author how to present his or her
information. This isn't a slippery slope we're on; we're *always*
interfering with authorial expression. The issue is how much.

>        b) the criteria are reasonably applicable to all websites in general

Oh?

How about Web applications? All-multimedia sites? All-Flash sites?

How well has this blandishment actually been *tested*? It hasn't, I
assume.


>        c) are machine or HHIR testable.

HHIR stands for...?

I warn against reliance on machine-testing as a goal. It took months to
explain to the Working Group that colour need not be machine-testable.
Untold hours of wild goose chasing ensued before the Group finally
accepted that, as with the correctness of an alt text, that issue must
rely on human judgement, inasmuch as it pertains to one participant's
cherished term, contrast. It is my impression that WAI wants to fit as
many square-peg testing criteria it can into the round hole of *machine*
testability.

I remind WAI that Web sites are forums for human creativity. They are not
credit-card numbers with check digits. A wide range of criteria can and
will always be testable only through human interpretation.

>    Category 3 = additional criteria that go beyond Category 1 and 2
>
>       that authors may want to consider if they want to make their sites
>
>       accessible or more usable to people with all or particular types of
>
>       disability.

If you present it that way, everyone will ignore Category 3. They're
obviously frills.

> - Some guidelines will have NO Category 1 items under them and are so
> marked.

which proves the folly of the approach. Category 1 is the basic level. How
can a guideline have no basic implementation? (If that is the case, why is
such a guideline its own category? Why isn't it collapsed onto another
guideline?)

> - The conformance would take the form of the familiar A, AA, and AAA.

or, more accurately, the confusing A/AA/AAA terminology.

> And, it solves most of the problems we have been facing:
>
>   - WCAG 2.0 would for the first time look like an evolution of WCAG 1.0

Why is that in any way even remotely desirable?

>   - The familiar A, AA, AAA would be there.

They're an albatross we need removed from our necks.

>   - Most of the WCAG 1.0 will match up with WCAG 2.0 except where the
>     working group believes that they no longer should.

Are we starting with a clean slate or not? Are we building a new house, or
whitewashing the old one?

>   - The guidelines would look more like 1.0 but would be generic.

The former is not a requirement for the latter.

>   - Transition would be easier to understand.

No, it wouldn't. WCAG 1.0 is already virtually incomprehensible to
non-insiders. A WCAG 2.0 that's even more complicated but is grafted onto
the structure of 1.0 is not going to be easier to understand.

--

  Joe Clark  |  joeclark@joeclark.org
  Author, _Building Accessible Websites_
  <http://joeclark.org/access/> | <http://joeclark.org/book/>
Received on Thursday, 6 November 2003 20:40:01 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:47:26 GMT