[Coco] FORTH for OS-9

Brett Gordon beretta42 at gmail.com
Mon Feb 29 02:24:43 EST 2016


Hmmm.. did anyone try Forth09?  I haven't, but I seriously studied
it's code.  Like basic09, it was build for modularity and memory
reuse.  It has a rare feature of any of the forths I've looked at.
One could specify a cost for each defined word, and specify a
threshold during the compiling state.  Based on that threshold, the
compiler either compiles in a jsr to the word, or directly compiles in
its action into the new word.  AKA, you could compile for speed or
code size, or somewhere in-between.  It's branching primitives where
just macro renaming of our 6809 friends bra,brn,beq,bpl.... It allows
for 2 vocabularies and target compiling areas, for compiling
stand-alone forth applications.  I really can't see this being any
slower than compiler C code.  (specially some of the garbage spewed
out by gcc.  I wonder how it would fair to cmoc code?)


Modern-day 6809 Forth efforts:

Jim O'Keefe: Working on a CoCo3/DECB port of the ultra standard FIG
forth 6809 spec sheet. (ASM)
Stephen M. Pereira: Working on CoCo2/DECB port of the ultra standard
FIG forth 6809 spec sheet. (ASM)
Me: I've done token based, VM based, and direct threaded forths for
the 6809. (ASM,FORTH,BFC,C)
David Given: "Hayes Tests" compliant ANS forth for Fuzix.
(AWK/GCC..yes, *AWK* :) )
Neal Crook: native Brad's R.'s famous Camel Forth for Multicomp. (ASM?)
"stickyak": Ninth Forth for os9. (GO/ASM)

I would dare say we have a small gaggle of active Forth programmers in
the community.  This is good to see.

Point of forth confusion:  There's two eforths. one for the coco, and
the more famous one by Dr. Ting.  Any one who cross-compiles should
read his work and the eforth source. It's great stuff.

forth go and m* !

-- 
Brett M. Gordon,
beretta42 at gmail.com


More information about the Coco mailing list