One idea using John's point about a library of domain
ontologies compared to the trouble at the upper ontology level: (01)
JS>Other groups have developed their own upper levels, and nobody
who has an upper level shows any interest in adopting anyone
else's. If Cyc hasn't accomplished this task in 23 years,
what is the evidence that another 23 years with more millions
JS>of dollars would make much, if any improvement? (02)
If one were to begin anew to urgently develop a "common foundation ontology "
then one might consider mapping the existing upper/foundational ontologies to
see what is common. At least that sounds as useful as getting some agreement
without understanding what is common in these. John's content seems to be that
you won't find large, obvious commonality. (03)
Gary Berg-Cross
SOCoP Executive Secretary
Potomac MD
________________________________ (04)
From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx on behalf of John F. Sowa
Sent: Wed 4/2/2008 10:52 AM
To: [ontolog-forum]
Subject: Re: [ontolog-forum] What is "understanding" (05)
Pat, (06)
I strongly endorse the idea of developing a library of domain
ontologies that have proved to be useful, along the lines that
are being proposed for the OOR. I believe that such an OOR
would be a valuable step that could eventually evolve into
something more complete, as further practical experience
accumulates and more R & D is done. (07)
But the following task, depending on how you interpret it,
is trivial, impossible, or in some vague intermediate stage: (08)
PC> That is why I think that getting some critical mass of users
> to agree on a common foundation ontology is an urgent issue. (09)
The impossible task is to develop a formally defined upper level
with every term specified by necessary and sufficient conditions.
That goal has been attempted repeatedly since Aristotle. Kant
tried and failed. Various 20th century philosophers, such as
Carnap, Goodman, and others, tried and failed. Cyc has been the
biggest attempt, but after 23 years and multimillions of dollars
of investment, mostly by the US government, it still cannot
support applications that can pay for its continuing R & D. (010)
Other groups have developed their own upper levels, and nobody
who has an upper level shows any interest in adopting anyone
else's. If Cyc hasn't accomplished this task in 23 years,
what is the evidence that another 23 years with more millions
of dollars would make much, if any improvement? (011)
At the other extreme is what lexicographers have been doing
successfully for the past several centuries: develop lexicons
and terminologies with loose definitions. Many broad-coverage
resources and specialized terminologies have been available
for years. The most widely used is WordNet. Aren't the 2000
words of Longman's defining vocabulary already in WordNet?
If any are missing, they could be added very quickly. (012)
As far as urgency goes, we have no agreement on what kind
of agreement is required. Cyc has been trying to sell their
approach to government and industry since 1984, some major
corporations (including Microsoft, for a couple of years
in the early '90s) bought into it, but none of them have
found much, if anything useful to do with it. (013)
What evidence do you have that any other group would be
able to do anything better, faster, or cheaper than Cyc? (014)
If your only complaint is that Cyc is not open source,
then you could suggest that the gov't buy Cyc and donate
it to the world. That would be cheaper than starting
a totally new program. (015)
John (016)
_________________________________________________________________
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 (017)
<<winmail.dat>>
_________________________________________________________________
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 (01)
|