Lists Home |
Date Index |
- From: "Bullard, Claude L (Len)" <firstname.lastname@example.org>
- To: "Sean B. Palmer" <email@example.com>, xml-dev <firstname.lastname@example.org>
- Date: Fri, 22 Dec 2000 08:46:17 -0600
I mean given a text, can one build an ontology
from it or match it to an existing ontology?
In other words, what does one do with an
ontology or knowledge level description?
One thing is to use it to establish
further deeper agreement about some
domain of discourse. I identified one
(a Request For a Proposal) and suggest
that each discourse by use of the ontology
may lead naturally and formally to the
next logical discourse. This is negotiation.
We can't entirely put away natural language.
In fact, we'd be hard pressed to give a
closed definition for "natural language".
If you can create an RDF from the web pages,
you can do it from other texts. The formality
and structure of the texts make a difference
of degree. Those definitions as resources
are in fact, natural language accompanied
by axiomatic tests.
It isn't impractical. Consider it in the
framework of UDDI defined services. A service
layer below discovery is interpretation and
negotiation in a cycle that outputs a Quote.
Then another round of negotiation to create
a Contract that will include Terms and Conditions.
This is followed by execution of said contract
which will include parallel operations, tests,
and so forth until the project
(outermost process of a nested process
BTW: UDDI provides elements with
key attributes (serviceKey, businessKey,
tModelKey, etc) and these can be used
to get details about services. The URL
gets you the entry point. At that point,
once you get past the surface descriptions,
you engage the local business protocol.
Businesses have been negotitating the
steps in such protocols since Rome ruled
Carthage. It is SOP.
Intergraph Public Safety
Ekam sat.h, Vipraah bahudhaa vadanti.
Daamyata. Datta. Dayadhvam.h
From: Sean B. Palmer [mailto:email@example.com]
> Could an ontology be used to service an engine
> that analyzes a text and determines the intent of
> the sender with regards to the text and what
> actions would meet that intent?
I don't think so. Note that TimBL in his XML2000 presentation said that the
Semantic Web did *not* mean the semantics of natural language. Of ocurse, I
may be completely misinterpreting what you are saying there! When you mean
analyze a text, did you mean analyze data?