Alaric B Snell <······@alaric-snell.com> writes:
> > Lisp is also a good language to write yourself. It's probably harder
> > than a simple Forth with system words written in C, but not much.
>
> I agree! The hardest bit, IMHO, is garbage collection. But YMMV.
Anecdote: one day I decided to write a Lisp implementation and was
merrily typing in the C code for it when the famous hacker and Lisp
guru Richard Stallman happened to walk by and look at my screen. He
asked what I was doing and I told him. He asked me, "Why are you
writing a Lisp interpreter? There are lots of those already". I
replied "everybody ought to write a Lisp interpreter" and expected
some retort from him (there were more important things that I really
needed to be doing instead). Instead, he thought about it for a few
seconds and seemed to decide that I'd given a good answer. He then
said "you ought to write it in assembler", and went on his way.
I decided he was absolutely right and that what I was doing was
pointless. I dropped it and did something else. I eventually did
implement Lisp in C as an extension language for something else I was
doing, but I never got around to writing Lisp in assembler. Maybe
someday still, but it would feel like pure retrocomputing now (it
almost was back when this anecdote happened).
Anyone ever written Lisp in Forth?
Paul Rubin <·············@NOSPAM.invalid> writes Re: Writing your own Forth
[..]
> Anyone ever written Lisp in Forth?
There are at least four documented attempts at writing "some kind of Lisp"
in "some kind of Forth". I think the problem with this is that you need
somebody who is very good in both languages. IMHO, these people have
much better things to do, or quit when they feel the essence of language
A has been caught in their implementation of it in language B.
Martin Tracy's "A Forth Lisp" in DR. DOBBS TOOLBOOK of FORTH II, Chpt 28
David Sharp's "Handling Linear Lists in Linear Memory" (LISP.txt in the
Taygeta archives), build on Martin Tracy's work.
lisp.frt (ANSified, with examples) in the iForth distribution, using David
Sharp's Lisp.txt.
lisp.xxx in kForth (?).
-marcel
> > Anyone ever written Lisp in Forth?
I've written a R3RS Scheme in Forth. It was for a Harris Semiconductor
processor called the RTX4000, a 32-bit Forth engine with 128-word data
and return stacks with over/underflow interrupts, and could be
externally microcoded. It was designed by Phil ·······@CMU, same guy
that did their RTX2000, and was originally designed as a graph
reduction engine (similar to the Haskell Tagless Spineless VM, though
several years earlier I think, and in hardware. Koopman describes it
in his dissertation). There were a small number of microcoded
instructions to help out (tagged add & subtract, tagged dispatch,
tagged car, cdr, set-car!, set-cdr!, structure element read and write,
some GC operations). Speed was a bit slower than the early Ivory
processors, depending on what you were doing -- subroutine calls were
*very* fast as might be expected for a Forth processor/Graph Reduction
engine, but it didn't have a FP processor, and the initial GC system
was a simple stop-and-copy collector. Harris sold their semiconductor
business and the company that bought the RTX line wasn't interested in
the RTX4000, so it never made it into production.
I've also got a fair chunk of a Forth system written in Lisp, just to
be contrary, based on FIG Forth, for even more contrariness points :-)
Dropped out of college back in the late '80's to help write a
Postscript interpreter in Forth (actually a dialect called, creatively
enough, "Fifth", which sort of a cross between Forth and Smalltalk),
and have always had a soft spot for the language. Fifth was the only
dialect of Forth that I've ever found suitable for large projects.
Every now and then I get back to the Forth-in-Lisp project, but I'm not
in any hurry.
Never seen a real lisp system in a generally-available Forth, only
list-handling systems based on reference counting. (My scheme system
was specific to the RTX4000 processor, and its microcode engine, and
its bizarre dialect of Forth).
Is there anything on the web where I could read more about your scheme
written in Forth, or your Forth in lisp?
How difficult a job was it to implement a scheme in Forth?
-----------
Michael Parker <·············@earthlink.net> wrote in message
> I've written a R3RS Scheme in Forth.
<snip>
> I've also got a fair chunk of a Forth system written in Lisp, just to
<snip>