Re: Test document draft

On Wed, 2007-03-14 at 09:18 -0400, Chimezie Ogbuji wrote:
> Below is a compilation of 'issues' (talking points, etc.) I came across 
> while organizing the tests (most of these have a @@ indication):
> 
> - No working reference to the 'public' CVSROOT for the test suite

That presumes there is a public CVS repository for the test suite.
I have not set one up; not yet, anyway. The only public repository
is the hg repository, which is currently only updated manually. 
 http://www.w3.org/2001/sw/grddl-wg/td/testlist1#test_scm
 -> http://homer.w3.org/~connolly/projects/grddl-tests/

>    (couldn't find grddl-wg anywhere in :pserver:anonymous@dev.w3.org:/sources/public)
> - Atom/Turtle test case has 'informative' status despite WG consensus

that's a simple matter of updating the testlistN file to cite the record
of the WG decision to approve the test. I don't have a clear memory of
approving it yet.

>    on the expected behavior, demonstrated support by at least 3 implemenations.  It was meant to 'close' an issue that has yet to be
>    addressed 'normatively' (#output-formats) and is relevant to recent LC cycle commentary
> - sq1a is supported by at least two implementations but is broken (shouldn't be empty)

yes, that's broken.

> - sq2 nominates an older version of sq1t.xsl (not sure why)
> - What is the status of the HTTP header mechanism?

we resolved 28 Feb to take it out and postpone the issue
http://www.w3.org/2004/01/rdxh/spec#issue-http-header-links
-> http://www.w3.org/2007/02/28-grddl-wg-minutes.html#item05

> - More security tests would be useful.  See Jena GRDDL test area [1]
> - Two properties were suggested in an earlier John-l patch [2] to capture (formally) the relationship between alternative test outputs:

How are those properties consumed? Data that isn't consumed rots.

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
D3C2 887B 0F92 6005 C541  0875 0F91 96DE 6E52 C29E