Re: reification test case

>On Mon, 4 Feb 2002, Patrick Stickler wrote:
>
>>  On 2002-02-04 17:23, "ext Dan Connolly" <connolly@w3.org> wrote:
>>
>>  > On Mon, 2002-02-04 at 07:12, Jeremy Carroll wrote:
>>  >> My vote: no.
>>  >
>>  > I vote yes.
>>  >
>>  > This is what "triple" means, after all, no?
>>  > if x=xx, y=yy, z=zz, then (x,y,z)=(xx,yy,zz), no?
>>
>>  But a bNode of type rdf:Statement is not a triple, it
>>  is the reification of a triple to which can be added
>>  additional knowledge such as authority, source, scope,
>>  etc.
>
>shall we say it is a 'description of' a triple? (avoiding the
>term 'reification' wherever possible strikes me as a useful strategy,
>at least while we're discussing rdf:Statement...)
>
>compare this to a 'description of' a Book, or a person, or any other type
>of thing whose instances might be described using bNodes in an RDF graph.
>In each such case the properties we attach to the bNode correspond to
>properties of the specific individual thing (some book, some person, some
>triple...) described.
>
>>  Whether two reification bNodes describe the same triple
>>  does not necessarily mean that other properties ascribed
>>  to each of those bNodes individually apply to all bNodes
>>  reifying the same triple.
>
>Trying this again swapping 'triple' for 'person':
>
>	whether two [reification] bNodes describe the same Person
>	does not necessarily mean that other propeties ascribed
>	to each of those bNodes individually apply to all bNodes
>	[reifying/describing] the same person.
>
>I'm not sure this works.
>
>Could you give me an example where we have two bNodes describing
>one-and-the-same Person and there are properties on one bNode that
>wouldn't be appropriately attachable to the other bNode.

Well, how would we KNOW that two bnodes describe the same person? If 
we could say that in RDF ,then there would be an obviously valid 
inference from

_:x sameAs _:y
_:x foo baz .

to

_:y foo baz .

But we can't say it, so there are no semantic constraints on RDF 
interpretations that could justify it, so its not valid in RDF.

>Or is there someone special about describing triples that makes
>it importantly different from describing people or books?

Yes, in fact. Nothing to do with the above, but in the case of 
triples, both the triple and its description can be in the same 
graph. That introduces a whole set of new issues. In the case of 
people, it would be like having portraits in the graph, and then 
having to worry about which rdf triples were about each picture. You 
can write as many more triples as you want, but none of that is going 
to get you to the right picture unless you provide some kind of link 
to pictures. How is the bnode in a  reification 'linked to' the 
triple that is being described? Does that question even make sense?

Pat


-- 
---------------------------------------------------------------------
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, 4 February 2002 18:59:44 UTC