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

Re: Revised model theory

From: <jos.deroo.jd@belgium.agfa.com>
Date: Fri, 27 Jul 2001 15:55:54 +0100
To: phayes@ai.uwf.edu
Cc: bwm@hplb.hpl.hp.com, w3c-rdfcore-wg@w3.org
Message-Id: <OFFA294FB0.86A07784-ON41256A96.00514689@bayer-ag.com>

[still not recovered from a very deep impression]

[...]
> 5. Skolemisation
>
> If we replace an anonNode by a uriref, the result is not entailed by the
original
> document. In fact, in general, if E' contains any urirefs not contained in E
then
> E cannot entail E'.  For example, the document containing the single triple
> <foo> <baz> _:xxx
> obviously bears some relation to the triple
> <foo> <baz> <bar>
> but it does not entail it, since the interpretation
> {1,2}/1->{<1,1>}/{foo->1,baz->1,bar->2}
> satisfies the first document but not the second. (Notice this interpretation
> assigns a value to something - <bar> - outside the vocabulary of the first
> document, which is how the entailment fails.)
>
> However, the second triple entails the first document, since any
interpretation
> which satisfies <foo> <bar> <baz> can be used to make the first document true
by
> restricting it to the smaller vocabulary and assigning the interpretation of
> <bar> to the anonNode to provide a suitable I[A] mapping.

i guess this is called "existential introduction", no?

> More significantly, any RDF expression which is entailed by the second triple
> and does not contain <bar> is also entailed by the first document.

why is that significant?

[...]

--
Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/
Received on Friday, 27 July 2001 09:58:06 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:38:18 EDT