Path: utzoo!attcan!uunet!zaphod.mps.ohio-state.edu!unix.cis.pitt.edu!pitt!willett!ForthNet
From: ForthNet@willett.pgh.pa.us (ForthNet articles from GEnie)
Newsgroups: comp.lang.forth
Subject: Atari Forth's
Message-ID: <2102.UUL1.3#5129@willett.pgh.pa.us>
Date: 13 Dec 90 12:34:37 GMT
Organization: String, Scotch tape, and Paperclips.  (in Pgh, PA)
Lines: 43

Category 1,  Topic 43
Message 9         Tue Dec 11, 1990
B.RODRIGUEZ2 [Brad]          at 22:48 EST
 
In reply to David Douthitt (rat@madnix.UUCP), re. SNOBOL functions implemented
in Forth...

> Can I get a gander at those when your done?

The first, very rough draft was done about a year ago, and reported in two
sections, at Rochester '89 and FORML '89.  See "Strings, Associative Access,
and Memory Allocation" by Dr. N. Solntseff and yours truly, Journal of Forth
Application & Research Vol. 6 No. 2, and "Pattern Matching in Forth" by yours
truly, in the 1989 FORML Conference Proceedings.

This version is rather heavily tailored for the 80x86; e.g., most of the
memory allocator is in CODE definitions.  (A weakness of mine.) Also, there is
at least one known bug in the pattern matching word set (the grouping
operators don't work right).  Finally, it's all written in a fig-Forth
derivative.  So it's not directly usable by others, which is why it hasn't
been posted yet.

If you're really determined, I'd be happy to send you the complete source
code, and (if I can find it) a disk copy of my thesis on the project -- about
200 pages, double-spaced.

> Phew!  2M+?  Bunga bunga.  Does it function in 1M?

Actually, it functions in less than 192K.  The complete package uses about 6K
of program space in my 64K Forth segment (simple 16-bit model).  It reserves
two more 64K segments for data: one for dynamically-allocated memory, and one
for the hash table.  (In SNOBOL, _everything_ gets hashed.)

One of the big reasons for the switch to the 68000 is that I want to write a
pedagogical version of the memory system, entirely in high- level.  The code
will be cleaner, simpler, and more portable if I can do linear addressing
within the Forth dictionary space using a 32-bit Forth.

- Brad

-----
This message came from GEnie via willett through a semi-automated process.
Report problems to: dwp@willett.pgh.pa.us or uunet!willett!dwp