W3C home > Mailing lists > Public > semantic-web@w3.org > August 2007

Re: on the semantics of RDF statements

From: Jeremy Carroll <jjc@hpl.hp.com>
Date: Tue, 07 Aug 2007 10:45:13 +0100
Message-ID: <46B83F29.8080609@hpl.hp.com>
To: K-fe bom <u9x3n_15so@hotmail.com>
CC: semantic-web@w3.org



K-fe bom wrote:
> 
> Most of the examples I see on RDF seem to assume data modeling semantics 
> equivalent to OO or database models of data. What if I model triplets with
> Noun Verb Object
> as subject predicate object respectively? Would I be infringing some 
> semantics of RDF and setting myself up for problems?
> 
> Thank you
> Gustavo Frederico
> 
> 

I think that's quite a common pattern. A problem comes when the Verb 
takes not only a subject and an object but a second object. e.g.

Jeremy gives Gustavo some advice.

This often ends up as three (or more triples).

e.g.

A)
eg:Jeremy eg:gives _:a .
_:a rdf:type eg:Gift .
_:a eg:recipient eg:Gustavo .
_:a eg:gift _:b .
_:b rdf:type eg:Advice.

Personally I then find the differing treatment of the subject as 
inelegant, and have a stylistic preference for:


B)
_:a eg:donor eg:Jeremy .
_:a rdf:type eg:Gift .
_:a eg:recipient eg:Gustavo .
_:a eg:gift _:b .
_:b rdf:type eg:Advice.

instead, where we are in the OO style of describing something, a gift.

Linguistically this is more akin to a deep structure rather than the 
surface structure; which may suggest that unary and binary verbs could 
also be modelled in the same way for consistency.

Given B) the following allows us to derive A) and other variants 
corresponding to

Gustavo receives some advice from Jeremy

Some advice is given to Gustavo by Jeremy

eg:donor owl:inverseOf eg:gives .
eg:recipient owl:inverseOf eg:receives .
eg:gift owl:inverseOf eg:isGivenBy.

Jeremy




-- 
Hewlett-Packard Limited
registered Office: Cain Road, Bracknell, Berks RG12 1HN
Registered No: 690597 England
Received on Tuesday, 7 August 2007 09:45:33 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 21:45:17 GMT