Dear Matthew, David, and Leo, (01)
MW
>> By names of types I mean things like pump and valve. Is that
>> what you mean? (02)
DP
> Yes, that's what I mean. There is nothing shared between the
> two ontologies at all. (03)
Just look at the very widely used Schema.org. Their hierarchy
is growing very large, but all the definitions are in English. (04)
Depending on your point of view, you could call it either
an underspecified ontology or a systematic terminology. (05)
In fact, some ontologies that had been specified in OWL, such
as GoodRelations, have been contributed to Schema.org. But most
of those ontologies use only a subset of OWL, their detailed
definitions are mostly stated in English comments, and they
have very few "upper level" commitments. (06)
DP
> We have not seen a case yet where two upper ontologies were useful,
> but I guess that is a possibility. (07)
MW
> I'm assuming you happen to have two bits of ontology that are
> pre-existing and that you want to make use of, that happen
> to have different upper level ontologies. I haven't come across
> an example either, but I guess it could happen. (08)
Again, look at Schema.org. Its upper level is very underspecified,
but it's widely used. Users who share information expressed in
the categories of Schema.org make no assumptions about a more
detailed upper level. Most of them probably have no upper level. (09)
By the way, Guha is Google's chief designer for Schema.org,
and he had been the associate director of Cyc -- until he and
Lenat parted ways in 1995. (010)
It's also significant that Guha wrote his PhD dissertation on
microtheories. He wrote that at Stanford in 1991, while he
was working at Cyc. His thesis adviser was John McCarthy, and
Ed Feigenbaum was also on his committee. Some implications: (011)
1. Guha certainly understands the theory, practice, pitfalls,
and applications (such as they are) of large ontologies. (012)
2. He developed his version of microtheories under the influence
of McCarthy's theory of contexts. He also *implemented* the
software for specifying and reasoning with the microtheories. (013)
3. I don't know the details about the Guha-Lenat split. But Guha
has publicly said that the lack of successful applications
(after the first 10 years of Cyc) led to his disillusionment
about the value of large formal ontologies. (014)
4. After leaving Cyc, Guha developed the MetaContent Facility (MCF)
at Apple and worked with Tim Bray at Netscape to turn it into RDF. (015)
5. MCF and RDF have a much, much weaker logic than CycL -- primarily
because Guha realized that ordinary humans did not have sufficient
training to use a more complex logic effectively. (016)
6. But Guha did not give up on logic, since he and Pat Hayes
specified LBase (Logic Base) for RDF and other SW logics.
The semantics of LBase, by the way, is the same as the core
semantics of Common Logic, which Pat was also developing. (017)
For further discussion and URLs of the original documents about
these issues, see http://www.jfsowa.com/ikl . (018)
That web page includes the URL of a talk by Guha about Schema.org:
http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2011_12_01 (019)
Leo
> The A-direct folks are generally ontologists/metaphysicians/some
> semanticists, The B-direct folks are generally logicians/some
> semanticists. B folks describe/represent what the A-folks say. (020)
I sympathize with your comments. But I would avoid making any
generalizations about metaphysicians, logicians, and scientists.
Note that Peirce and Whitehead were pioneers in logic, who also
published research in physics and developed major ontologies.
For a summary, see http://www.jfsowa.com/pubs/signproc.htm (021)
Observation: Schema.org is extremely important for applied ontology.
The Wikipedia terms are also very important, but they're even less
formal than Schema.org. Other resources of similar scope will
undoubtedly evolve as time goes by. (022)
Recommendation: If you can't fight 'em, join 'em. If you want
people to use your formal ontologies, map them to Schema.org
and/or Wikipedia and/or other widely used terminologies. (023)
John (024)
_________________________________________________________________
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 (025)
|