In response to my request for a public domain numerical analysis and statistics package in LISP, Mike Wirth (···@rice.edu) writes:
How about XLISP-STAT from Luke Tierney at U. of Minn? Here are pointers to
resources:
Buy the book: "LISP-STAT: An Object-Oriented Environment for Statistical
Computing and Dynamic Graphics"
Luke Tierney, Wiley, 1990.
email the author: ····@umnstat.stat.umn.edu
ftp the code: host umnstat.stat.umn.edu in directory pub/xlispstat
(versions for Mac, Unix and Amiga are available).
Most of the code is written in C and uses XLISP as a front-end.
I have downloaded XLISP-STAT but I haven't had the opportunity to evaluate its performance since it is written partially in XLISP (as indicated above) and we run Lucid Common LISP. Documentation comes in the form of a well-written manual/tutorial which assumes little prior experience with the LISP language.
Much of XLISP-STAT's functionality is implemented through XLISP's message-based object system. My guess is that a port to Common LISP (and CLOS) would be fairly straightforward. However, XLISP-STAT is purely a statistical package and my original post requested a numerical analysis package with emphasis on curve fitting techniques--i.e., something along the lines of what you might encounter in an undergrad introductory course in numerical methods for computing. Apparently, most LISP programmer's are using f
oreign function calls to existing math packages implemented in C or FORTRAN. If anyone out there knows of a numerical analysis package *in LISP* I am still interested.
Meanwhile, if you need statistical functionality in LISP I reccommend that you check out XLISP-STAT. As Bluto said "Don't cost nuthin."
--
*
Len Charest, Jr. ·······@ai-cyclops.jpl.nasa.gov
JPL Artificial Intelligence Group
*