RE: Need compelling story on the value of ontologies in fusing location-based data

> But Danny, the OWL ontology cannot tell Roger's app what to do
> with the data
> that it classifies as "inside" (what it "means" to be "inside").  His app
> has to know that beforehand.

It's getting a bit tangled, so just to requote Roger :

> It would be *extremely* cool if, by consulting an OWL Location ontology,
> an application could recognize that the location specified in document 2
> is located *within* the bounding box specified in document 1.

What I'm trying to say is that although parts (the calculations) of this are
outside of the immediate domain of OWL, it is still plausible that the whole
thing could still benefit from using OWL. For a start it would be possible
to identify the entities involved : doc2, loc, doc1, box and the important
relationships : e.g. loc specified in doc2, loc inside box. (this latter
relationship may not be true).

There are a lot of things in the meaning of 'inside' which can't be
expressed, but one of the most important ones can : if something is inside
it can't be outside can be, e.g. by saying the set of all inside things is
disjoint from the set of all outside things.

That the calculations are out of the immediate domain of OWL isn't really
the last word there - they could be expressed as a web service, which would
get called as needed to answer questions about the domain. The location and
purpose of the service could be specified using RDF, and it's necessity
could be inferred from rules applied to the data presented.

What is also possible given such a setup is that the inference engine could
be used to draw further conclusions, or prompt actions based whether or not
(loc inside box) is true.

Like I said earlier, the good bit is that (on the OWL/RDF side) this is all
built from standard, reusable parts. The application wiring (and state) can
all be expressed declaratively.

Cheers,
Danny.

Received on Friday, 27 June 2003 07:58:53 UTC