Prolog Glue Language?

[Jon Fernquest]

chris wright wrote:
>I've written a native python logic programming framework if that's helpful.
>It has quite good unification performance and I'm using it as the basis for
>some reactive planning stuff with BDI agents that I've implemented in
>Python.....if you really need to have the performance.

Sound's great.
I'd like to try it out.
Is it available somewhere?
Does it use SLD resolution, like Prolog?

Having a good logic/ai/nlp toolset available would be great, e.g.:

1. term unification (w/ and w/o occurs check)
2. dag/graph unification
3. lambda expression reduction
4. skolemization
5. reduction to predicate logic normal forms
6. a strips like planner
7. finite state tools
8. earley parser...etc

A lot of these tools are available at the
CMU AI repository in Lisp form.
(I rewrote several of them in Perl but I began
to feel like I was reinventing the wheel.)

Tim Peter's wrote:
>...or try to get the world interested in a brand new
>PrologScript language <0.9 wink>.

Python is perfect given my goals. Why reinvent the wheel?
I just want to make as many high-powered tools available
in the application domain I'm working in, CALL (Computer Aided Language
Learning) so language teachers (who seem to be limited to Javascript
right now, ugh, or the Mac's HyperCard which is great but not portable,
non-standard, and really doesn't have that much to glue together anyway)
can whip together something of their own creation from trustworthy
components...gui,linguistic/nlp,multimedia,web,database,dtp...etc..

Everything in the current CALL repositories of software is platform
specific (usually Mac or PC) and monolithic take it or leave it,
not take what you want and leave what you don't,
and tends to focus on repetitive grammar drills...language is a lot
more than grammar...e.g. listening comprehension, research on the
structure of conversation:

http://www.cs.umd.edu/users/traum/
http://www.cogsci.ed.ac.uk/~poesio/
http://www.ling.gu.se/~cooper/
http://www.informatik.uni-hamburg.de/WSV/hp/schilder-english.html

could be integrated with audio to better explain what's
happening in dialogue. (It's pretty difficult to explain
in writing how intonation patterns convey meaning in conversation.)

IMHO
Ousterhout's paradigm of the glue language applied with Python
could raise the productivity of CALL and create yet another
corp of non-programmer programmers,
(language teacher programmers (scary?))
that Guido van Rossum envisions in his essay
"Computer Programming for Everyone"
http://www.python.org/doc/essays/cp4e.html

Cheers,

Jon Fernquest
bayinnaung at hotmail.com