NetNews Usenet Archive 1992 #16/NN_1992_16.iso/spool/comp/lang/forth/2855

Internet Message Format | 1992-07-25 | 5.9 KB

Path: sparky!uunet!gatech!pitt!willett!ForthNet
From: [e-mail] (ForthNet articles from GEnie)
Newsgroups: comp.lang.forth
Subject: An Open ANSI Forth Implementation
Message-ID: <3939.UUL1.3#[e-mail]>
Date: 25 Jul 92 12:45:49 GMT
Organization: EIEI-U
Lines: 110

Category 10,  Topic 41
Message 10        Fri Jul 24, 1992
D.RUFFER [Dennis]            at 19:34 EDT

Re: [e-mail]

 > Could you present evidence for your premise that commercial
 > success of a language comes from vendor profit?
 >
 > Looking at the languages currently in commercial use, I see that
 > it's the other way round: first the language became popular
 > (usually in academia), then it became commercially used and then
 > vendors could make a profit (in spite of free or low-cost
 > academic and hobbyist implementations).  This is true for Pascal,
 > C, and in some sense also for FORTRAN.

Anton, Mitch will probably respond also, but I think you are _very_ far off
base with your assumption.  I have not followed Pascal's development and can't
get to a library at the moment, but I think that is the _only_ language that
fits your remark.  C was _only_ written so that Unix could be developed _and_
it was pushed heavily by AT&T which is about the biggest vendor I could
imagine.  They gave it to universities so that they could force it to be
popular. That would also work for Forth, if we have an AT&T behind it, but we
don't so it hasn't.  Does anyone even remember how FORTRAN started and are
they still alive today? ;)  I doubt there even were university courses for
software back in them days and in those days, it was most likely a computer
manufacturer (probably IBM) who made the languages so that they could sell the
computer.

I'm sorry, but if you are going to throw up a straw man opinion as an argument
against vendor contributions to languages, at least do a little bit of
research to get your facts straight first.  In Forth's case, it probably would
not even exist if Forth, Inc. had not been (and still is today) doing
everything it can to promote its use. IMHO that has been the case for _every_
language that did not come directly out of the universities themselves.  Of
course, now days, about the _only_ thing that academia will even look at is
something they have developed themselves.  Talk about the NIH syndrome,
they've got it _really_ bad.

Re: [e-mail]

 > I would probably not have bought any, had I not been introduced
 > to Forth via freebie Forths on each platform.  (The first hit is
 > always free I guess. :-)

You are absoultely correct.  The PD Forths are _essential_ to get new users
interested in the language.  We absolutely do need a good PD ANS Forth.
However, to think that the success of FIG Forth could be repeated in this day
and age is a little bit of a pipe dream I think.  First, the world is a whole
lot larger than it was back in 1978 and there are a _lot_ more experts out
there who could care less about what new wiz bang Forth is developed.  Most of
them will upgrade their own versions to be ANS compliant.  They might use a PD
one as an example, but I doubt any will throw away the versions they have
poured all of their efforts into over the years.

 > I'd rather see all that effort go into developing tools and
 > libraries for one widely available ANSI Forth, than to have a
 > dozens of widely available ANSI Forths which all have the same
 > minimum level of functionality.

While I certainly share your desire, I still think that it is a dream.  Forth
is just too extensible for one version to _ever_ fit everyone's needs.

 > Since there doesn't seem to be any chance that any type of
 > internal standardization will happen, the next best thing, (for
 > me anyway,) if I'm interested in portability, is to rely on a
 > common, widely-available, platform existing on all my target
 > machines.

Again, I completely agree.  Everyone should pick a system that is:
        1) Widely available,
        2) Well supported and
        3) Runs on all the machines they will be using.

I just doubt that you are going to find those qualities in the Public Domain
anymore.

 > Finally,  vendors will always be able to tout their
 > 'direct-jsr-threaded-inline-peephole-optimized' Forth languages
 > with the nifty development systems as being superior to the
 > public-domain version. This is wehere the real selling point to
 > Forth systems is anyway,  the Forth engine itself may as well be
 > given away for free.

The "engine" (or at least a "model" of it) has always been public domain and
will continue to be so for ever, I imagine.  However, this does not mean that
_everyone_ has to, or wants to, use the generic version for their
applications.  Typically, it is _not_ the vendors that are touting their
"tricked up systems", but the expert users who are always playing with the
internals.  Typically, once a vendor has a model to work with, they will
continue using it until something very tramatic comes along to make them
change.  Shifting ground is not something that customers like very much and
_every_ successful vendor will do anything they can to protect their
customer's investments.

So, there is no one Forth engine today, and IMHO ANS Forth is certainly not
going to change that situation in the future.  It will merely make it slightly
easier to move an application from one vendor's system to another.  I say
vendor here, because the PD systems also do have "vendors" behind them.  Laxen
& Perry for F83, Frank Seargent for Pygmy, Ting and SVFIG for eForth, Tom
Zimmer for F-PC and yes, FIG for FIG Forth.  The _only_ difference is the
amount of support that these vendors provide after you have "purchased" their
systems.

Just who is a vendor anyway?   {B-{)>   DaR
-----
This message came from GEnie via willett.  You *cannot* reply to the author
using e-mail.  Please post a follow-up article, or use any instructions
the author may have included (USMail addresses, telephone #, etc.).
Report problems to: [e-mail]