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

Re: SUO: RE: An article on the pitfalls of metadata




Rich,

I stated more qualifications to that point:

> Yes, in fact John Sowa recently opined that WordNet, for all
> its lack of axioms, appears to be the only ontology that
> has an empty top level.  All the others force a specific
> top level before providing multiple interpretations at
> lower levels.  As I remember John's comment (it would be
> useful if someone can find it in the archives), he felt the
> imposition of a specific top level set is a widely practiced
> mistake that ontologists make.

WordNet does have a top level, which nobody uses, and everybody
who has an ontology replaces with their own pet top level.

I didn't say that a top level is bad, but that it should not
be imposed from the top down.  Instead, it should be derived
from the bottom up (or better, by an iterative method that
is subject to many kinds of tests against reality and
applications to reality).

John Sowa