Re: Guidelines Implementation

There are up to four kinds of things we are tracking relationships among:

a: computable test conditions

b: manually evaluated conditions

c: conclusions inferred from the evidence.  some of these are satisfaction
relationships to [guidlines and checkpoints] defined elsewhere [primarily
WCAG].
Some of these are belief-state grayscale values regarding those
satisfaction propositions, such as "the satisfaction of checkpoint 1.3 is
cast in doubt."

d: corrective actions to take [edit document, query user, ...]

The logic flow is a [ * b ] =logic=> c [ =heuristic=> d ]

Notation:

[ foo ] denotes that the clause "foo" is optional and may or may not be
present
* denotes logic 'and'
=logic=> denotes "implies by logical decuction or tautology"
=heuristic=> denotes "selects by preponderance of expert and statistical
evidence"

At this point, if I were designing a table of these relationships, I would
start with a: at the left column and enumerate them that way unless it came
time to split rows for different combinations of a: and b: where some
values of b: led to a unique c: or d:

It is possible to start with checkpoints and subdivide, listing a+b
combinations, but somehow that is now how I would list them.

Al

At 03:05 PM 5/26/99 +0200, Daniel Dardailler wrote:
>
>> Actually, it was my understanding as well that your "'can/cannot/maybe can'
>> be automated" checklist (evalauto) would continue to sit on the w3c server
>> and that any comments that someone from the group made would be crossed
>> linked to the relevant section of the archive list (both ways). For
example,
>> I might send an email to the list...
>
>The modified checklist (serving as start page into our eval technique
>pages) doesn't have to sit on w3.org while it's being edited (depends
>on how easy it is to give write access to our site)
>
>I'm currently discussing with Chris Ridpath to see whether or not he
>feels ok being the editor of this resource. If he does, then maybe the 
>resource can sit on utoronto for a while.
>
>Eventually, it will have to sit on w3.org, in ER space. 
>

Received on Wednesday, 26 May 1999 13:01:01 UTC