magik@wlcrjs.UUCP (Ben Liberman) (07/11/85)
Here are the replys to my request.
Subject: Re: advice wanted on forth machine
Newsgroups: net.lang.forth
--------------------
From: ihnp4!uw-beaver!entropy!fetrow (David Fetrow)
Buy anything with standard disk I/O. That's the beauty of it,
any fig-forth micro acts a lot like any fig-forth Cray. (I'm
assuming fig-forth, I'm not all that familiar with the "Real"
forths).
-Dave Fetrow
APL forever
------------------------
From: ihnp4!cadovax!keithd (Keith Doyle)
After porting several versions of Forth to various processors, I find that
the 68000 is far superior to the other 16 bit processors out there for
supporting Forth. This is primarily because of the fact that the 68000
architecture makes it easy to work with 2 stacks at the same time, while
the 86 family and others can make it rather difficult (read - extra overhead).
In the 8-bit world, it seems that the 6809 is one of the better choices, as
it has a 2 stack architecture, (and it has integer multiply). This might
imply the Coco (Radio Shack). I've never owned a Radio Shack computer, so
don't take this as a particular endorsement, but I've gotten the impression
that they produce a fairly popular 6809 machine.
However, if you just want to get some exposure, and are considering buying a
new machine for this purpose, I would suggest that you might determine some
other useful criteria for selecting a potential machine (other than Forth
suitability) as you can find Forth compilers available for just about any
machine for relatively few bucks, and you might want to use the machine for
other things as well.
I've a lot of experience in CP/M-80 machines and I picked up the Forth system
produced by Laboratory Microsystems (Santa Monica, CA, - see Dr. Dobbs or Byte)
and found that they have an excellent package, though perhaps not the cheapest.
I'm not sure of the present price of their package for such machines, but it
may be as much as $100-150 range. There are a variety of public domain Forth's
for such machines, but obviously you get little documentation and no support.
LMI produces Forth's for various machines, and they are cheaper than the
average 'C' compiler, so if you are not totally strapped for a buget, then
I would consider them as a excellent source for the software.
Dr. Dobbs has a Forth issue every September, a good source of Forth examples,
and if you call Mountain View Press (check any handy BYTE) you can join
FIG and get their publication (Forth Dimensions) which is also chock full
of Forth ideas and techniques. Also, Mountain View Press has quite a few
versions of Forth and various publications available, they are a potential
source of related information.
Keith Doyle
# {ucbvax,ihnp4,decvax}!trwrb!cadovax!keithd
------------------------
From: ihnp4!tektronix!tekgvs!keithe
If you end up getting a Commodore 64 ('cause it's *cheap*) then
I'd recommend SuperForth 64 from Parsec as the Forth implementation
to get for it. C64 Forth (from somebody else) is not as good,
particularly in the disk I/O.
keith ericson at Tek Labs
------------------------
From: ihnp4!vax135!phri!hplabs!nsc!voder!scgvaxd!kontron!steve.ttyi5 (Steve McIntosh)
Soon, (probably by the time you read this) the Atari 520 ST will be on
the shelves. For $800 you get a half meg of ram, an 8 mhz 68000, a half
meg floppy and a hi-res monochrome monitor.
There are already 2 forth systems available for the machine (check out
the latest issue of Forth Demensions or wait for the ads to show up in
BYTE).
I would suggest avoiding Forth-83 implementations, and stick with a
Forth-79 implementation.
It is a sure thing that a public domain version of forth will be available
for the ST within a few months.
------------------------
--
-----------------------------------------
Ben Liberman {ihnp4|ihldt}!wlcrjs!magik