Re: SUO: What the marketplace needs from us.
Adam,
I agree with you that SUMO would fall into
the "neat" category because it uses logic for
representing ontologies. I think that we both
agree that is a good idea.
> I think most folks would call SUMO neat in its
> adherence to a formal semantics without appeal
> to any procedural code, and its use of FOL.
The comparison of "primitives" vs. "space cadets"
was used to characterize the kind of people who
write code that gets used vs. those who write
papers with abstruse levels of formalism that
never gets used in any application.
I would characterize Robert Kent as a space
cadet, whose theories strike fear in the hearts
of most nonmathematicians. I would like to
encourage Robert to bring his ship closer
to earth by relating it to SUMO and OpenCyc
and working with people who can implement
a usable subset of the theory.
I would characterize the way that SUMO was
developed with a minimum of tools as using
a primitive style of coding, where the code
happens to be a logic-based language called
KIF. I would like to see appropriate tools
applied to SUMO to relate it OpenCyc and
other content sources in a systematic way.
What I believe we need is a balance of both
kinds of talents: logic-based representations
of content written by people who are addressing
practical applications with theory-based tools
designed by people who are developing systematic
methods for organizing and manipulating that
content.
The result is neither primitive nor space cadet,
but a judicious balance of the two.
John