Re: Directions for SUO: monolithicity and Cyc
At 7:45 PM -0800 12/3/01, Michael Uschold wrote:
>I'm afraid that I have been unable to get to any SUO mail lately, but
>I did have a peek at this message from Nicola - promoting a non-monolithic
>approach which emphasizes understanding and relating ontologies, rather than
>agreeing on a single one.
>
>If it turns out that this is agreed by a everyone, this could have a major
>shift in our emphasis and activity.
>
>There is another possibly very important development which I believe we should
>take into consideration. Doug Lenat recently announced that they will be
>giving away Cyc.
I believe that, if CYC becomes available, we *must* consider it, as
well as any other sufficiently well-recognized ontology.
>
>How does this affect the SUO effort? For those that advocate a non-monolithic
>approach, it Cyc will merely become one of those that need to be
>available, and related to other ontologies. For those advocating a
>monolithic one single SUO, it can have much greater import.
>
I agree.
>I belive that before much more energy goes into building our SUO, 'we' should
>instead do a systematic and thorough investigation of CYC, and
>answer the above question. The answer might be, there is no reason
>to build another
>UO. Alternatively, it should take the form of a thorough critique of Cyc
>indicating the shortcomings, and why it is insufficient (if it is) and how it
>needs to be augmented or replaced by a 'real' SUO..
... But in order to perform such systematic and thorough
investigation we need to choose some formal "tools", and we must
clarify our own ontological assumptions. For instance, in order to
claim that some CYC's generalizations are inconsistent, I need to
establish some analysis principles based on identity criteria...
Example: in Cyc, an organization is both a social entity and a group
(unless they changed something recently...). Since social
organizations and groups have incompatible identity criteria (social
organizations can change members, groups can't), then this
generalization is inconsistent (assuming we adopt such criteria, of
course).
>
>An alternative thread of activity can be as Nicola recommends, to spend time
>relating alternatives, and work towards a library of reference ontologies, one
>of which will be Cyc.
Taking Cyc as a whole "module" is not what I suggest; rather, we
should try to "steal" from Cyc as much as possible, adding more stuff
from our own if necessary.
-- Nicola
---------------------------------
Nicola Guarino
National Research Council phone: +39 O49 8295751
LADSEB-CNR fax: +39 O49 8295763
Corso Stati Uniti, 4 email: Nicola.Guarino@ladseb.pd.cnr.it
I-35127 Padova
Italy
http://www.ladseb.pd.cnr.it/infor/ontology/ontology.html
(***updated 22/2/2001 ***)