Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: SUO: single vs. multiple ontology standard: modularization?





Pat,

You have some very good points.   I tend to use "module" in this case in a
broader sense that I haven't really thought through.   WordWeb (a thesaurus
running in MS Windows based on WordNet) gives:

     Module:  A self-contained unit or item that is used in combination
     with other units

This may be a bit too broad, but I'm not going to try to narrow it now.

I think that in engineering and in everyday life we implicitly switch
between possibly incompatible "micro theories" to bring the most
appropriate level and perspective of abstraction to bear on a given
problem.  SUO should (in my view) provide for this kind of usage.  This
requires somewhat more planning and design than simply collecting an
aggregate of micro-theories.  The points you have raised can serve as an
outline for some of the requirements on a modularized SUO.

Easy for me to say.  Sorry, I don't have time to do anything concrete (yet).
               ^^^

John Velman


Pat Cassidy wrote --
>   In this line, the term "modularization" has
>been used, but I am unclear as to its intention.
>In programming, modules in general are not
>inconsistent with each other.  One simply doesn't
>have to use all of them.  Certainly the SUO
>can have "modules" that can be used or ignored at
>will, even though *none* are inconsistent with
>any other.  A manufacturing ontology won't
>usually need the biochemistry or medical "module",
>and if they are sufficiently independent, they may
>be excluded from a working manufacturing ontology
>and then added later if they are needed, without
>causing any inconsistencies.  This will still
>give us the interoperability that is one of
>the prime goals of having any standard.  I think
>that Adam's point about CYC can also be
>interpreted as meaning that the "microtheories"
>are mostly "modules" in this sense.
>    We will also need some mechanism to handle
>inconsistent theories, possible worlds, fictional
>contexts, etc.   But this is very different
>from the usual use of "modules" in programming.
>A "lattice of theories" seems to be the more
>appropriate term for this mechanism.
>    Are we all using the same terminology?
>
>    Pat Cassidy
>
[snip]