Re: SUO: Automated or Semiautomated Ontology Development
Bill,
The main point of my recent note is that Aristotle's paradigm is
a healthy strategy. In a nutshell, it is a recommendation to
look at the data instead of staring at your navel. For ontology,
that means that you should look at what people say about something
as a starting point for doing knowledge representation. Aristotle
did that without using computers or automated NLP, and human bodies
who write axioms today can do the same.
In fact, that is my first recommendation for anyone who is trying
to write axioms in any formalism -- KIF, CGs, SQL, or even C.
Before writing down anything in the formal language, first write
down a clear statement in your own native language of what the
axiom (or program) is supposed to say (or do).
It would be great if we had NLP programs that would automatically
translate such a statement into code, but it is a very healthy
(i.e., OntoClean) strategy for any programmer or ontologist to write
a clear NL specification before starting to write the code. (The
final result may require a complete rewrite of the spec, but having
one as a starting point is always a good way to go.)
And many years ago, Dijkstra made a similar recommenation (even though
he had no intention of doing any NLP): the first qualification that
he looked for in programmers is the ability to express themselves
clearly in their own native language. If they can't express themselves
in their own language, they're not going to be able to do so in a
computer language (or in any version of logic).
BA> Step up to the plate and write the code, and we'll see if it works.
The two talks I mentioned discussed a lot of code that has already been
written that does many of the necessary tasks:
http://www.jfsowa.com/talks/negotiat.htm
http://www.jfsowa.com/pubs/autotalk.htm
And I'll be writing more later to spell out more of the details with
more pointers to actual running programs.
John Sowa