Re: SUO: Re: Re: Variables, Of A Sort

["Robert E. Kent" <rekent@ontologos.org>]

John, and others,

In general, the "logical languages" that I referred to are mathematical
objects, much like groups, topological spaces, etc. In particular, they are
objects in a category of languages formed of languages and language
morphisms, much like groups and group homomorphisms form a category Grp,
topological spaces and continuous maps form a category Top, etc. These
"logical languages" represent the notion of an object language, whereas the
categorical framework in which they are embedded represents the notion of a
metalanguage. The whole approach is the idea of mathematics modeling logic.

These "logical languages", which represent the idea of logical language in
model theory, are close to, but not the same as, Chris's description
[http://ltsc.ieee.org/logs/suo/msg01792.html] of "sorted languages/logics."
They are definitely not AI languages. However, I believe they will prove
very useful in modeling an appropriate sorted logical AI language to express
the SUO and other ontologies. Such an appropriate AI language might take the
form of a "fully sorted KIF".

Robert

----- Original Message -----
From: "John F. Sowa" <sowa@bestweb.net>
To: "Robert E. Kent" <rekent@ontologos.org>; "Jon Awbrey"
<jawbrey@oakland.edu>; <standard-upper-ontology@ieee.org>
Sent: Monday, October 30, 2000 3:58 PM
Subject: Re: SUO: Re: Re: Variables, Of A Sort


>
> Robert,
>
> Designing a new language is very, very easy.  Designing a good
> language is very, very hard.  And even if you have done the
> very, very good work to design a very good language, getting
> anyone to pay any attention to it is extremely, extremely hard.
>
> Questions:  Why are you designing a new language?  Why do you
> think that anyone would bother to look at your design?  In what
> way is it better for what you want to do than KIF or any other
> AI language?
>
> If there are features in your language that cannot be adequately
> expressed in KIF, that would be an excellent topic for SUO list,
> since we are in the process of taking a good, hard look at KIF
> to decide whether or not to add, delete, or modify any of the
> features of KIF'98 to come up with SUO-KIF.
>
> Have you tried to translate your language to KIF?  If not, why
> not?  If so, which features were easy to map to KIF, and which
> were difficult or impossible?
>
> John Sowa