Re: proposed test like rdfms-seq-representation/Manifest.rdf#test002

>Pat says [1]:
>
>>  Full disclosure: there is a case which COULD have been a test case
>>  but isn't, which would have been affected if had been a test case,
>>  which is
>>
>  > { } entails { _:x rdf:type rdfs:ContainerMembershipProperty .}
>>
>>  which is not an valid LC2-entailment.  That is, the current semantics
>>  requires that some containermembership properties must exist, even if
>>  nobody talks about them.
>
>which is just tricky enough that you should probably add it as a test
>case.

Its mentioned in the change log of the semantics document.

>  My implementation would have to change to handle it, but it
>would be an easy change.  For some implementations, it could be a real
>pain.  (Queries like { ?x rdf:type ?y } will be unable to return
>complete answer sets.)

No, that will be OK.  There is one new kind of answer:

?y/rdfs:ContainerMembershipProperty

with no binding to ?x, ie Yes, something of that type exists and No, 
I don't know its name.

So I don't think that any implementation should find this (new) case 
very hard to handle, compared to some of the others. That query is 
pretty peculiar in any case (is anything in any class?)

Pat

>I'll consider no response a satisfactory response for this comment,
>given its timing.
>
>       -- sandro
>
>[1] http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2003Nov/0095.html


-- 
---------------------------------------------------------------------
IHMC	(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32501			(850)291 0667    cell
phayes@ihmc.us       http://www.ihmc.us/users/phayes

Received on Wednesday, 12 November 2003 15:56:11 UTC