Domains and ranges (was:Re: new version of semantic layering document)

>[...]
>
>>  >>  BTW, strengthening domain and range to IFF seems to me to be a really
>>  >>  bad idea, in general: it completely destroys the idea of conjunctive
>>  >>  range/domain assertions. We put the current semantics into RDFS in
>>  >>  order to conform to the DAML conjunctive semantics, so why do you
>>  >>  want to blow this out of the water in OWL?
>>  >
>>  >I don't understand why you are saying this?  How does making domain and
>>  >range IFF destroy the conjunctive semantics?
>>
>>  Because if superclasses of ranges are also ranges, then there is no
>>  information to be gained by asserting more about a range. All
>>  properties range over the universe. The whole point of allowing
>>  conjunctive semantics on ranges is so that one can accumulate
>>  information that allows one to pin down a range more precisely.
>
>Huh?
>
>If I say
>
>A:
>	ex:bar rdfs:subClassOf ex:baa .
>	ex:baz rdfs:subClassOf ex:bar .
>	ex:foo rdfs:range ex:bar .
>
>Then no matter whether ex:foo has an IFF reading or an ONLY IF reading
>any object of an ex:foo triple must belong to both ex:bar and ex:baa
>so
>
>B:
>	ex:bar rdfs:subClassOf ex:baa .
>	ex:baz rdfs:subClassOf ex:bar .
>	ex:foo rdfs:range ex:bar .
>	ex:a ex:foo ex:b .
>
>RDFS entails
>
>	ex:b rdf:type ex:bar .
>	ex:b rdf:type ex:baa .

Right, of course. Whch is why Jeremy's entailment is beside the point.

>The only difference is that under an IFF reading graph A RDFS entails
>
>	ex:foo rdfs:range ex:baa .

Right, and I don't think it should. Here's an analogy. If I'm in my 
house, then I'm in Pensacola. I live in Pensacola, true: but it would 
be a mistake to infer that my house *was* Pensacola.

>
>If I say instead
>
>C:
>	ex:bar rdfs:subClassOf ex:baa .
>	ex:baz rdfs:subClassOf ex:bar .
>	ex:foo rdfs:range ex:bar .
>	ex:foo rdfs:range ex:baz .
>
>then more information has definitely been gained, as C RDFS entails
>
>		ex:foo rdfs:range ex:baz .
>
>which was not an RDFS entailment of A.

Well, yes. But on your semantics that doesn't tell me very much, since

ex:foo rdfs:range rdf:Resource .

no matter what you tell me, so I can't use any of these range 
assertions to localise the range more exactly. If all we were 
inferring was rdf:type assertions then it wouldnt matter much, but 
ranges are used for more than that. You have suggested attaching 
datatyping information to ranges, for example; if all properties have 
the universe as a range then this becomes impossible.

>
>Further
>
>D:
>	ex:bar rdfs:subClassOf ex:baa .
>	ex:baz rdfs:subClassOf ex:bar .
>	ex:foo rdfs:range ex:bar .
>	ex:foo rdfs:range ex:baz .
>	ex:a ex:foo ex:b .
>
>entails
>
>	ex:b rdf:type ex:baz .
>
>which is not an RDFS entailment of  B.

Yes, but this is true under either interpretation of range, right? I 
fail to see what point you are making here.

Pat
(rest of response on another thread)
-- 
---------------------------------------------------------------------
IHMC					(850)434 8903   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola,  FL 32501			(850)202 4440   fax
phayes@ai.uwf.edu 
http://www.coginst.uwf.edu/~phayes

Received on Monday, 23 September 2002 22:17:42 UTC