W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > September 2001

Re: Reification in RDF MT (probably off-topic)

From: pat hayes <phayes@ai.uwf.edu>
Date: Mon, 10 Sep 2001 20:08:35 -0500
Message-Id: <v04210111b7c313344551@[205.160.76.173]>
To: Graham Klyne <GK@NineByNine.org>
Cc: w3c-rdfcore-wg@w3.org
>At 07:51 AM 9/7/01 -0700, pat hayes wrote:
>>>(Not strictly a WG topic for now, but I wanted to test my understanding...)
>>>
>>>In the 28-Aug MT draft, section 4, the reification conditions are 
>>>described as
>>>
>>>(V union VxVxV) is a subset of IR
>>><x,y> is in IEXT(I(rdf:subject)) iff for some a,b,c in V, x=<a b c> and y=a
>>><x,y> is in IEXT(I(rdf:predicate)) iff for some a,b,c in V, x=<a b 
>>>c> and y=b
>>><x,y> is in IEXT(I(rdf:object)) iff for some a,b,c in V, x=<a b c> and y=c
>>>x is in ICEXT(I(rdf:Statement)) iff for some a,b,c in V, x=<a b c>
>>>
>>>This seems to say that *every* 3-tuple in VxVxV is of type rdf:Statement.
>>
>>In VxVxV, yes. Not every triple in IRxIRxIR is, of course.
>>
>>>This appears to preclude the use of the 3-tuple for any other purpose.
>>
>>Er.... .not sure what 'use' means here. You can assert anything you 
>>like about the triple, since its in IR. But it is indeed of type 
>>rdf:Statement.
>
>It's probably a red herring.  But I was wondering what would happen 
>if one wanted, in future, to allow 3-tuples of resources in the 
>domain of interpretation for some purpose other than representing a 
>statement.  I find it difficult to imagine a useful node that 
>wouldn't be some part of some statement, hence be a member of V.

I have no trouble at all, eg how about a triple of real numbers, or 
of pieces of coal, or of plays by Shakespeare, or whatever. (They may 
not be nodes, but they are triples.) The things in IR are the 
entities being denoted, remember, not their URIs. Putting V and VxVxV 
in the domain just says that some of the expressions are counted as 
things that can be talked about by other expressions, but there are 
still a hell of a lot of other things that can be talked about and 
are not expressions, so are not of type rdf:Statement.

There can be triples in IR already. The proposed V-semantics for 
reification is just a way to be explicit about some of the triples 
being syntactic in nature, ie in the class rdf:Statement.

Pat

---------------------------------------------------------------------
(650)859 6569 w
(650)494 3973 h (until September)
phayes@ai.uwf.edu 
http://www.coginst.uwf.edu/~phayes
Received on Monday, 10 September 2001 21:07:07 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:39:41 EDT