[Top] [All Lists]

Re: [ontac-forum] Follow up question on Ontology, knowledge, languagecon

To: ONTAC-WG General Discussion <ontac-forum@xxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Sat, 08 Oct 2005 10:08:49 -0700
Message-id: <4347FD21.6060808@xxxxxxxxxxx>
Steve,    (01)

I agree:    (02)

 > Creating a unified upper ontology, without consensus of
 > the creators of the leading upper ontologies, is unlikely.    (03)

The Standard Upper Ontology WG has led to several projects
in which people have merged terms and axioms from other
ontologies, but the people who developed the original
ontologies have never accepted anyone else's merged
ontologies as suitable replacements for their own.    (04)

When two groups get together to talk, the developers in
each group often modify or expand their own ontologies,
but I have never seen any group voluntarily abandon their
own ontology in favor of anyone else's (except under
outside pressure, such as an offer to grant or withdraw
a large pot of cash).    (05)

The following point requires a lot of qualification:    (06)

 > The focus of Cyc, and much of the AI'sh/ontology work, has
 > been concentrated on finding ways to replicate human thought
 > which by its nature is a very "researchy" activity. In Cyc's
 > defense, their mission to create the world's first "thinking
 > machine", is a research activity.....    (07)

Lenat's ultimate goal has always been very grand, but at every
step along the way (since 1984), industry groups and gov't
agencies have been working with Cyc and trying to use its
ontology and other resources for practical applications.    (08)

 From the beginning, Cyc had industry partners that were paying
half a million dollars per year for the rights to use any or
all of the Cyc resources any way they liked.  In 1991, one of
the Cyc sponsors was Microsoft, and at that time I was working
at IBM.  I made a proposal that IBM should become a Cyc sponsor,
but IBM Research management said that they had better uses for
the money.  I couldn't argue with that point, and a couple of
years later, Microsoft also dropped their support.    (09)

So you can't blame Cyc's lack of commercial applications
on a lack of effort.  That is why I am very skeptical about
the idea of just building ontologies with the hope that
something useful will magically happen.    (010)

I agree with that point:    (011)

 > I'd like to see a bit of working backwards from some problems
 > the government wants solved to how ontology can help.    (012)

I also agree to a large extent with the talk that Kevin Hannon
presented on October 5:    (013)

http://colab.cim3.net/file/work/SICoP/ontac/meeting/2005-10-05/Hannon-Multiple_v_Single_Taxonomies_04.ppt    (014)

His point was that many small ontologies are more useful,
more flexible, and easier to build, deploy, and maintain
than one giant ontology.  There are, of course, many issues
to be discussed about how to relate those ontologies, but
I don't believe that a global ontology is either necessary
or sufficient to relate them or to make them interoperable.    (015)

In fact, I would say that interoperability *never* depends
on a global alignment of two or more ontologies, but *always*
on a local task-oriented agreement or a negotiation toward
an agreement that is sufficient for the task at hand.  But
that is a whole 'nuther topic, as they say.    (016)

John Sowa    (017)

Message Archives: http://colab.cim3.net/forum/ontac-forum/
To Post: mailto:ontac-forum@xxxxxxxxxxxxxx
Shared Files: http://colab.cim3.net/file/work/SICoP/ontac/
Community Wiki: 
http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/OntologyTaxonomyCoordinatingWG    (018)
<Prev in Thread] Current Thread [Next in Thread>