W3C home > Mailing lists > Public > www-webont-wg@w3.org > September 2002

RE: possible semantic bugs concerning domain and range

From: pat hayes <phayes@ai.uwf.edu>
Date: Mon, 23 Sep 2002 13:54:20 -0500
Message-Id: <p05111b38b9b512131f0a@[65.217.30.172]>
To: "Jeremy Carroll" <jjc@hplb.hpl.hp.com>
Cc: www-webont-wg@w3.org

>FYI
>
>this topic is also being aired in RDF Core where I produced the following
>test case:
>
>1:[[
>
>eg:prop rdfs:range eg:A .
>eg:A rdfs:subClassOf eg:B .
>
>entails
>
>eg:prop rdfs:range eg:B .
>]]
>
>(i.e. in english,
>every object of a eg:prop is an eg:A.
>every eg:A is an eg:B,
>thus
>every object of an eg:prop is an eg:B).

Thats not a fair transcription. You can get to *that* conclusion 
without changing the definition of range, just by using subClass 
reasoning. Your English conclusion here is best expressed by saying

eg:prop rdfs:range _:x .
_:x rdfs:subClassOf eg:B .

which of course does follow from your premis, by simple RDF generalization.

Your conclusion would be a fair rendering of "every object of an 
eg:prop is an eg:B and vice versa"

>
>The current RDF MT says this does not hold.
>Peter's OWL semantics says this does hold.

Right, and I think the current RDF MT has it right. If we adopt the 
stronger version then we can't conjoin range constraints, for 
example, since

eg:foo rdfs:range ex:A .
eg:foo rdfs:range ex:B .

would entail both
ex:A rdfs:subClassOf ex:B .
and
ex:B rdfs:subClassOf ex:A .

so that seems to me to be a non-starter.

Pat


>Jeremy
>
>Pat:
>>  Ah, it occurs to me that one of the bugs that Peter may have been
>>  referring to is my use of rdfs:range and rdfs:domain.  Peter
>>  apparently believes that the RDFS semantics for these are wrong and
>>  need correcting. However, I disagree, and do not propose to alter
>>  them in the RDFS MT. If OWL needs to use different notions then OWL
>>  should introduce and use owl:domain and owl:range rather than use the
>>  RDFS vocabulary. I would however suggest that the adoption of a
>>  different mechanism at such a basic level might be a decision which
>>  should be examined very carefully, as it has many repercussions (eg
>>  for datatyping mechanisms). I havn't seen any good arguments for it,
>  > which is why I simply used the RDFS notions in the document.
>  >


-- 
---------------------------------------------------------------------
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 14:54:13 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:52 GMT