On Wed, Jan 18, 2006 at 05:23:14PM -0500, Obrst, Leo wrote:
> Of course, an alternative is something like the Information Flow
> Framework meta-ontology of Bob Kent, which addresses (last I heard)
> only FOL level (though Bob wants to raise that) using Barwise &
> Seligman's information flow theory, itself an application of category
> theory. Ways to compare ontologies, syntactically, semantically, do
> the equivalent of the projection of two ontologies, etc. Later, used
> (both IFF and IFT) by Kalfoglou and Schorlemmer for ontology
> mapping/semantic integration in IF-Map.
>
> I know the FOL folks have a bit of heartburn with this (Chris), (01)
*urp* Pardon! :-) (02)
> because their usual question is: why all the rigamarole when FOL
> languages (like presumably IKL) can do what you want without all this
> category theory biz. I don't know the answer, but maybe IFT/IFF gives
> you better ontology-level constructs? (03)
I suspect that category theory might ultimately provide a more general
and elegant formal framework for semantic integration. My suspicion, in
a nutshell, is that for a lot of folks it is overkill. I wrote a paper
about this for the 2004 Dagstuhl Workshop on semantic integration, in
case anyone is interested in my take on the issue: (04)
http://drops.dagstuhl.de/opus/volltexte/2005/42/ (05)
Here's the first paragraph: (06)
The important work of Joseph Goguen ([6], [7]), Robert Kent ([8]), Marco
Schorlemmer and Yiannis Kalfoglou ([11], [12]), and others point the way
toward very promising general framework for characterizing of a variety
of concepts of ontology integration. Such high-level frameworks are
essential for the sort of theoretical foundation for semantic
integration needed to bring it to the level of a genuine engineering
discipline. At the same time, this work is done in the rarefied
theoretical air of category theory and channel theory, and therefore
assumes a certain amount of background knowledge that a lot of people
working in ontology, even at a fairly high theoretical level, lack. In
fact, however, while this work is far more abstract and, concomitantly,
far more general and far-reaching in its implications and applicability,
I believe some of the most basic insights beneath the idea of semantic
integration can be expressed in terms of basic first-order logic and
model theory. Moreover, I believe it is important to do so to provide
relatively simple, comparatively concrete accounts of integration that
can help to fix the basic ideas of the emerging theory for the broader
community of ontological engineers. The major purpose of this brief
paper, then, is provide a simple model of integration that remains
within the friendlier confines of first-order languages and their usual
classical semantics and logic. The model might also serve as a sort of
"bidirectional" test-bed for the higher-level theoreticians as well --
any virtues of the approach that are not reflected in the higher-level
theories can be appropriated by them, and any infelicities in the
approach can be corrected on general grounds provided by the theories. (07)
> By the way, sorry to continue another distracting thread: the problem
> is that these are all interesting threads but they really don't lead to
> work getting done. (08)
I fear this group is too big and contains too many people from a
diversity of backgrounds already committed to a diversity of approaches.
It does not augur at all well when what is essentially a quibble over
the advantages of "type" over "class" generates so much controversy. (09)
Chris Menzel (010)
_________________________________________________________________
Message Archives: http://colab.cim3.net/forum/ontac-dev/
To Post: mailto:ontac-dev@xxxxxxxxxxxxxx
Subscribe/Unsubscribe/Config: http://colab.cim3.net/mailman/listinfo/ontac-dev/
Shared Files: http://colab.cim3.net/file/work/SICoP/ontac/
Community Wiki:
http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/OntologyTaxonomyCoordinatingWG (011)
|