In an attempt to be constructive, here are some thoughts to hopefully
refocus the discussion in preparation for the ontology summit ... (01)
We want a definition of ontology that can address the following questions:
- how is Alice's ontology related to Bob's ontology? For example, can they
be used together or is there some conflict between them?
- can Alice reuse (and possibly extend) Bob's ontology for her domain? (02)
Proposed working definition:
An ontology includes a vocabulary together with a specification of the
meanings of the terms in the vocabulary. Furthermore, this specification
is definable in a language with a model theory and a machine-interpretable
syntax. (03)
Notes:
1. Model theory defines notions of truth assignment, entailment,
satisfiability, validity (04)
2. If we consider the specification of meaning to consist of a set of
sentences
in a formal language, then there is the unresolved problem of
distinguishing
an ontology from an arbitrary set of sentences in the language. (05)
3. This definition does not require that the ontology be specified in a
formal
language. Ontologies specified as taxonomies, Topic Maps, or similar
approaches
are included so long as the concepts are definable in a formal language.
For example, taxonomies are definable in monadic FOL, Topic Maps are
definable in Common Logic (or even RDF?), and Leo has suggested similar
ideas for thesauri. (06)
4. This definition does not require that all specifications of all
ontologies be
written in the same language. All we need is
- a formal characterization of relative expressiveness for logical languages
- a characterization of the minimally expressive logical language
required to
define the concepts in a specification that is not in a logical language. (07)
Again, we are looking for a definition of "ontology" that can support the
sharing and reuse of ontologies. (08)
5. If an ontology is specified in two different languages with different
expressiveness,
then the relationships between the two languages should be sufficient to
show
that the two specifications are logically equivalent. If the
specifications are in fact
not logically equivalent, it would perhaps be more accurate to say that
they
correspond to two different (though related) ontologies. (09)
- michael (010)
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2007/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2007
Community Portal: http://ontolog.cim3.net/ (011)
|