Re: Test cases for parseType mis-spelling errors

On Sat, Oct 20, 2001 at 11:35:44AM +0100, Brian McBride wrote:
> 
> Can I suggest that test cases which do not involve WG issues are something you 
> can just collect, don't need reviewing on a case by case basis and can be 
> reviewed as part of the process of publishing WD's.  Less work, all round, yes?

Seems like reviewing the non-issue related test cases in pieces or 
in one batch would take the same amount of time.

> The charter states as deliverables:
> 
>    * publish a set of machine-processable test cases corresponding to
>      technical issues addressed by the WG
>
> I think this would be a great thing to do, but I 
> guess I have to act as charter policeman a bit here.

Well, if you're going to start playing charter cop, then I don't
believe the charter explictly lists as deliverables test automation 
or automatic triple generation :-).  

You may also want to consider a threshold for the number of WG members 
that need to approve a test case before it gets labeled as Approved 
in the TC doc.

> Are you envisaging just a few extra test cases here and there, or something 
> requiring more significant effort?

I currently have no plans for creating new tests.  I do, however, 
agree with the sentiments Graham expresed in:

 http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Oct/0429.html

and wouldn't categorically deny a test case just because it
doesn't represent an issue in the issue tracking doc.

Art
---

Received on Wednesday, 24 October 2001 16:17:50 UTC