Re: EARL Testcase

Yes, we need the binding.

Can you suggest a method of binding the tool specific test to the standard
test suite?

Chris

----- Original Message ----- 
From: "Shadi Abou-Zahra" <shadi@w3.org>
To: "'Chris Ridpath'" <chris.ridpath@utoronto.ca>; "'WAI ER IG List'"
<w3c-wai-er-ig@w3.org>
Sent: Thursday, April 15, 2004 1:21 PM
Subject: RE: EARL Testcase


> hi chris, hi all,
>
> this is indeed my problem when trying to construct an "EARL Wrapper
> Tool". how does one aggregate data from different sources?
>
> my thought is that the best solution would be to have the earl:testcase
> element bind a tool specific testid (for example
> http://vendor-a.org/tool/tests/#check-for-alt) to a testcase of a test
> suite which all sources (tools or humans) in the collection are testing
> against. in this case it would be for example
> http://www.w3.org/TR/WCAG10/#tech-text-equivalent **.
>
> when aggregating (or exchanging) the results, one could compare the
> assertions conducted on the testcase (marked by ** in the paragraph
> above) and make more comprehensive statements (for example two tools
> can't tell pass/fail, one tool says it failed, etc). the real tests
> conducted (for example http://vendor-b.org/tool/tests/#check-alt-sanity)
> become secondary but interesting algorithms could be based on using the
> earl:mode and the earl:assertedBy.
>
> any thoughts on this?
>
> regards,
>   shadi
>

Received on Thursday, 15 April 2004 14:08:59 UTC