Re: Hidden triples and self-description

On Sun, 14 Sep 2003, Mark Baker wrote:
>
> On Sun, Sep 14, 2003 at 05:38:53AM -0400, Bill de hÓra wrote:
> > Mark Baker wrote:
> >
> > > I think that's different.  All parties may know RDF, but those that know
> > > the rules of entailment will extract additional triples from the same
> > > message.
> >
> > It's not different. Knowing RDF means knowing its entailments, there
> > isn't anything else to 'know'.
>
> Oops, sorry, should have said "those that know the *RDFS* rules of
> entailment".  Follow?  Hence my RDFS media type suggestion, as it
> communicates that there are different entailments at work (i.e. makes
> the message more self-descriptive).
>
Consider the following two scenarios:

When two RDF only applications communicate, the sender will transfer all
relevant triples, as it does not know anything about entailment. Same for
the receiver, it does not expect anything more, and thus will understand
exactly what is said.

When two RDFS (or OWL, or whatever) aware applications communicate,
both know about entailment. The sender does not have to send any entailed
triples, and thus can keep the message short and to the point. The
receiver can extract any "hidden" triples through the entailment rules,
and thus again understand exactly what is said.

The case you described appears only when you cross the layer boundaries
between RDFS aware applications and plain RDF only applications. As long
as you are aware of this crossing, you can write your applications to
prepare their messages accordingly. A new mimetypes could solve this issue
in a general way - but then you would need a new mimetype for OWL too, and
any possible layer crossing! I am not sure if this really is so common a
problem to justify all the trouble.

Regards,
Karsten

Received on Monday, 15 September 2003 05:34:34 UTC