Lists Home |
Date Index |
Roger L. Costello wrote:
>I think it is clear that except for trivial, academic cases RDF Schema
>and OWL do not have the robustness to capture the dynamically changing
>nature of real-world semantics. To do so, we must go beyond these
I think it is clear that either we are not using the same definition for
the term "ontology" or else you don't understand how ontologies have
been developed and deployed for real world applications.
More real world examples, the SNOMED and GALEN medical ontologies, as
I've said, licensed for use by both the U.S. and U.K. governments for
healthcare applications (healthcare is a roughly 1.5 trillion/year
industry in the U.S.).
>I have compiled a somewhat random (chaotic) list of statements which I
>feel expresses much of what has been discussed:
>- Ontology languages such as RDF Schema and OWL provide the ability to
>*statically* capture semantic relationships.
>- Semantics is constantly changing. All of life is constantly
>changing. In fact, change is the only constant.
You need to understand how RDF and OWL define "semantics" in a very
precise although highly mathematical fashion -- I'm not sure how you
arrive at your conclusion that they only capture *static* semantics. A
central idea behind the type of semantics that is being used here, is
that the mappings are to sets of *possible* worlds, not a single
specific world. The idea is that when we converse, as long as we have
some shared understanding of the words we are both using, we can have a
conversation. This doesn't require that we both use *exactly* the same
definition, merely that our differing definitions contain enough overlap
or context so that we might disambiguate what we are saying.
For example let's use the term "4 dollars" ... now this might mean
canadian dollars or american dollars and what you might be able to
purchase with "4 dollars" depends on a whole host of spatial and
temporal variables. This has no real bearing upon our ability to develop
an ontology which contains terms such as "cost" which involve values
such as "dollars" (admittedly it is a good idea to specify "U.S.
dollars" vs. "Canadian dollars").
>Semantics as graphs Semantics as high-complexity
Err... ontologies are being used for high-complexity science, for
example molecular biology/genomics/proteomics e.g. the NCI Ontology
(developed in OWL), SNOMED, GALEN as well as Astrophysics e.g.
ontologies developed for the NSF/NASA NVO project, etc. etc.