Re: rdfc25May.n3 testcase

jos.deroo.jd@belgium.agfa.com wrote:
> 
> I thought to do a straightforward test namely
> asserting the facts contained in
> http://www.w3.org/2000/11/mr76/rdfc25May.n3
> and then querying with that same
> http://www.w3.org/2000/11/mr76/rdfc25May.n3
> which should of course always succeed.
> But it didn't succeed...

Hmm... let me try that with cwm... hmm...
it's not working; I'm not sure why...

> After some tracing I found
>   n:_23 [ mr:from who:mdean; a mr:Absent ]; # didn't actually call his name, did we?
>   n:_23 [ mr:from who:guha; a mr:Absent ]; # didn't actually call his name, did we?
> giving the trouble (2 times n:_23) which I think
> is indeed an inconsistency. Right?

Not by itself, no.

If you combine it with an axiom that n:_23 (and
all ContainerMembershipProperties) and
mr:from are UniqueProperties and who:mdean
and who:guha are different, then you get a contradiction.


> That wasn't all, because there are lots of
> mr:result properties further on as well.

Right; in the general case, it's perfectly
fine to have several values for the same
property of the same subject:
	:s :p :o1.
	:s :p :o2.
	:s :p :o3.

in fact, there's a short-hand syntax for it in N3:

	:s :p :o1, :o2, :o3.

(but I don't expect to use any shorthands for
test case expected results; I intend to add
some sort of --simple output that just
writes one statement per line.)

> So I am in doubt now...

It's not clear what you're in doubt about.
Is it about the use of N3 for test case output?
Or about the idea of graph matching on
expected results in general?

> --
> Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/
> 
> ps KIF reminded be to use log:forSome in a query

-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Wednesday, 30 May 2001 11:00:01 UTC