Peter and Michael, (01)
I agree with Martin Hebb's observations. The only thing I would change
is the title. My recommendation would be (02)
Useful ontologies vs. research ontologies (03)
Unlike the researchers who propose elegant formal ontologies that
nobody uses, Martin has been designing simple ontologies and the
supporting tools that application developers can actually use. (04)
PY
> I think [Martin's] identification of the intrinsic differences between
> "traditional" ontologies and web ontologies is an important insight,
> and am therefore, sharing the links to Martin's talk[1] in the hope that
> this may generate some discussion and, hopefully, help bridge the
> understanding between the communities that profess in them.
>
> [1] see: http://vimeo.com/51152934 (05)
Thanks for posting the URL. I recommend the talk. (06)
From the talk by MH
> "The degree of detail and expressiveness in an ontology is inversely
> correlated to the achievable community size" (07)
MB
> I subscribe to this. At Web scale, OWL may be too expressive instead
> of too restrictive. (08)
The words 'expressive' and 'restrictive' are not antonyms. Programming
languages, for example, are highly expressive and totally unrestricted.
I would say that OWL hits a "sour spot" in knowledge representation. (09)
The point Martin emphasized is that the upper level ontology should
be very underspecified (almost no axioms). I enthusiastically agree.
That property is shared by Martin's Good Relations and most of the
published ontologies on the WWW (which don't use any expressive power
beyond what Aristotle specified over two millennia ago). (010)
That's a lesson Lenat & Guha learned over 20 years ago when they were
developing Cyc. Instead of a large monolithic ontology, they split the
ontology in a very general, but underspecified core and an open-ended
collection of microtheories that contained detailed the axioms needed
for each branch of applications. (011)
But an underspecified ontology, by itself, can't support an application.
You need either (a) a richer logic or (b) a programming language. When
I say that OWL is a "sour spot", I mean that it's too rich to specify
useful ontologies, but not rich enough to implement applications. (012)
That's one more reason why I say that Tim B-L was on the right track
in his proposal of 2000. He had much more experience with implementing
practical systems and better insight than the DL researchers into what
was needed. (013)
John (014)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: 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 join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (015)
|