Re: TEST: proposal for test case structure

Sorry Dan I had missed your message:

>> - use manifest file to list tests and their status
>(where is the RDF core manifest, btw? I don't recall seeing it.)

Files named Manifest.rdf

>> 1: Directories of directories of test cases.
>> I suggest the following:
>> is a directory.
>deep hierarchies are bad.

Why? personally I quite like them ...
(Genuine question, perhaps not very relevant)

but ...

> I'd rather just

> I'd rather just

OK, I'll go with these.

(I think I'll omit the F in Fxxx e.g.
for the feature InverseFunctionalProperty I will use a directory named

> where xxx refers to a section of the language description document.

I think it will be easier to navigate if xxx (usually) refers to some item
in the owl namespace.

DanC (on [in]consistency tests):
> I would need to see some examples of these to be
> convinced this is worth doing.

For all the test types I imagine lazy evaluation ... i.e. we might have a
sketch of our test types now, but we really only have a test type when we
have an instance of such a test type!

>> 5.2 Illegal document test.
>> An illegal document test consists of one document.
>> bad001.rdf
>or even
>	bad001.txt

Yes, but the only examples we have to date are good RDF, just improper use
of owl namespace (non-language features).

>> An issue that appears is that the current test case area is already
>> with files that do not adhere to these proposed conventions.
> - don't pay much attention to the files that don't conform to
>	our conventions.

Oh yes, that works.
Particularly if we generate index.html files (or whatever they are) which
expose only the "good" files. This also works for the N3 stuff etc.


Received on Friday, 26 July 2002 02:11:32 UTC