Paola, I have been trolling the internet to see if I can discover
what recent work you were referring to in our recent email exchange.
Your use of the phrase 'collaborative and modular ontologies' was
very helpful. I hadn't realized from the wording used in your Wiki
document that this was the subfield you were meaning to refer to. (01)
Did you meant to refer to the kind of work illustrated for example by (02)
[1] http://wonderweb.semanticweb.org/deliverables/documents/D21.pdf
[2] http://www.cs.iastate.edu/~honavar/Papers/BaoISWC2006.pdf (03)
which presume that ontologies are packaged into 'remote' and 'local'
parts, with an implicit scoping of concept names to the 'local'
ontology? As the references above note, this whole approach is
inspired by the success of such 'modular' thinking in the design of
distributed software systems. If so, I have some comments. (04)
I think this work is interesting, but I am not (yet) persuaded by the
arguments given that it is either desirable or necessary, or that the
analogy with software development is really accurate. I feel that the
'open publication' of ontological content on the Web, and the use of
a single global name space of IRIs to indicate concepts, is far more
likely to lead to useful interoperation between ontologies and
distributed composing of useful content, by concept re-use. This is
the really important aspect of the semantic Web, IMO: it removes the
need to think of 'modules' in this way. One simply re-uses concepts
from other ontologies in ones ontology, importing if necessary, but
not being obliged to (contrary to what Stuckenschmidt presumes in
[1], by the way). I think that the whole idea of a 'local semantics'
is in fact a strategic mistake, and misses an important sense in
which the Semantic Web global-name-space and open-publication
paradigms represent a quantum jump in potential ontological utility.
The workshop description for the 2006 Athens meeting
(http://www.cild.iastate.edu/events/womo.html ) says the central
presumption clearly (my emphasis): (05)
"... Because no single ontology can meet the needs of all users under
every conceivable scenario, the ontology that meets the needs of a
user or a group of users needs to be assembled from several
independently developed ontology modules. Thus, in realistic
applications, it is often desirable to logically integrate different
ontologies, wholly or in part, into a single, reconciled ontology.
Ideally, one would expect the individual ontologies to be developed
as independently as possible from the rest, ..." (06)
Notice the presumption, almost indeed stated as a requirement, that
the ontologies be developed independently from one another. Why??
This is exactly how they should not be developed, IMO: and the
Semantic Web paradigm, for the first time, gives us an opportunity to
overcome this 'independence' assumption. The whole point of
publishing ontologies is that other ontologies do not have to be
developed in isolation, independently, but that one ontology can
re-use concepts from other ontologies. Elsewhere we have referred to
this as a 'distributed syndication' model of content development. (In
some ways it is similar to the open source model of software
development rather than the commercial managed models: by publishing
"code" openly, you obviate the need to constantly define and
re-define APIs.) (07)
Anyway, If I at least have got your intended reference point right,
we can disagree about something substantive rather than who is or is
not in or out of a loop :-) (08)
(I do however agree with a point often made in this literature, that
simply importing an entire ontology is too coarse a tool, and we need
better ways to refer to 'pieces' of a large ontology.) (09)
Pat
--
---------------------------------------------------------------------
IHMC (850)434 8903 or (650)494 3973 home
40 South Alcaniz St. (850)202 4416 office
Pensacola (850)202 4440 fax
FL 32502 (850)291 0667 cell
phayesAT-SIGNihmc.us http://www.ihmc.us/users/phayes (010)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (011)
|