Re: ERT WG: Agenda for teleconference on Tuesday 5 April 2005

Probable regrets. 2am is just a bit unlikely for me (although I may manage  

On Mon, 04 Apr 2005 21:38:19 +1000, Shadi Abou-Zahra <> wrote:

> 2. EARL Processing Model
>   - Let's discuss how we can expand the PM
>   <>

A suggestion: The nature of each test should determine what kind of  
locator information is most useful for that test. This assumes that we  
define different kinds of locator. That way we can describe the relevant  
ones for each test without describing anything more about the test (and  
falling further down the slippery slope of a test definition language  
before we need to).

> 3. Conformance to EARL
>   - Chris posted some sample of EARL based on WCAG tests
>   - Assuming the EARL is correct, is that what we need?
>   <>

It's not correct, and that is a real problem in determining if it is what  
we need.

It shows a handful of possible locator possibilities, and doesn't make it  
clear if they should all be used together, or some should and some should  
not, or what.

It doesn't cover the case of an assertion made as the result of an  
inference (mode="heuristic"), where I think we shold enable the evidence  
to be provided.

I think there is a question of whether we need a confidence rating, given  
the interoperability problems with it.

If we are going to do this we should note that there are a number of  
possible forms that the output can take, since it is RDF. (I think there  
are a number of reasons wy we should not lead people to assume that it  
will also conform to some XML schema. The short version is that it breaks  
interoperability with RDF).

But it is certainly on the right track.



Charles McCathieNevile                      Fundacion Sidar   +61 409 134 136

Received on Monday, 4 April 2005 15:10:17 UTC