Re: EARL in Page Valet

On Mon, 4 Feb 2002, Charles McCathieNevile wrote:

> Do you think it would be useful if we had a "warning" property, or do you
> think it is beter to make pass/fail assertions and use confidence levels? It
> seems to me that there might be a lot of value  in "warning" properties
> being seperate since having confidence levels for real assertions is helpful
> to be able to deal with conflicting reports.

I'm in two minds on this one.  At the moment, an accessibility warning
with a "Low" confidence level is effectively just a warning.  But the
purpose it serves is to say "this tool can't tell whether XYZ violates
accessibility".  A warning is the right thing to present to a human
reader, but an automatic system needs to know there are/aren't
any such issues.

I think for my purposes I'd prefer to stick to the confidence levels
scheme, noting that a Low Confidence assertion from Page Valet can be
nullified by a contradictory High Confidence assertion from
elsewhere, but should not be silently nullified by default.

Note that the currently assigned confidence levels for different
tests/messages are rather arbitrary, and subject to review.

> Alternatively, it might just lead to things getting lost.
>
> Another approach to this would be to work on flagging conformance to
> techniques for WCAG, and building profiles of those techniques which enable
> you to make in inference that a particular collecction of techniques being
> implemented is equivalent to satisfying a checkpoint.

Erm - that sounds like an altogether different project.  If it ever
happens, I'll be looking to support it as well as, not instead of,
WCAG 1.0.

-- 
Nick Kew

Site Valet - the mark of Quality on the Web.
<URL:http://valet.webthing.com/>

Received on Monday, 4 February 2002 17:01:46 UTC