Path: utzoo!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!uunet!samsung!usc!sdsu!ucsd!ames!amdahl!pacbell!sactoh0!tree!stever
From: stever@tree.uucp (Steve Rudek)
Newsgroups: comp.lang.forth
Subject: Re: Declining Forth popularity.
Summary: comments on your comments
Keywords: jforth, jforth, jforth, jforth, ...
Message-ID: <1989Dec24.192309.3889@tree.uucp>
Date: 24 Dec 89 19:23:09 GMT
References: <8912160503.AA25980@jade.berkeley.edu> <15140@well.UUCP>
Organization: TREE BBS (916)349-0385 Sacramento, Ca
Lines: 109

In article <15140@well.UUCP>, jax@well.UUCP (Jack J. Woehr) writes:
> In article <1989Dec18.185612.8335@tree.uucp> I said:
> >The VENDORS have made almost zero contribution to the evolution of
> >language standards but appear to have been dragged every step of the way.
> 
> 	Not true. Forth, Inc. still sells the best commercially available
> multiuser Forth operating system; Harvard Softwarks pioneered full-mem-model
> 80386 Forth; LMI UR/Forth hopped into the OS/2 market before most software
> houses knew a DLL from a dildo; Delta Research maintains what is still
> the premiere 68000 Forth, JForth, which incorporates most of your
> suggested improvements and much much more, and has done so since 1986.
Note I said "evolution of STANDARDS".

> 	Hypertext is a feature, no-one is going to argue about that
> in a Forth system. Neat feature, very helpful, has no more impact
> on Forth itself than the color of your keyboard.

I guess it may be a bit unrealistic to hope for a Forth "hypertext"
standard.  Still, it became evident to me after seeing VIEW in F83 that
Forth, more than any other language, can make use of a hypertext-like
language documentation facility.  Hypertext is Forth's best hope
for a resurgence in popular favor.  I said earlier that Forth's learning
curve was too steep.  That, of course, is not quite correct.  The learning
curve may actually be LESS steep than with many other languages.  It's just
long.  Long, long, very, very, long.  And it's mainly because of the huge
number of words involved in the language.  And the fact that you pretty much
MUST know ALL the words and how they interrelate before you can program
anything significant in Forth.  Forth is sort of like the Mandarin Chinese
of the programming languages--I understand that the number of words in F-PC
version 3.5 approaches 3,000.  Mandarin Chinese doubtless gives you
exquisite control of nuances, also, but I keep hearing how large it is and
how hard it is to master. 

A hypertext "standard" for Forth is probably too much to hope for.  But it
is what Forth very much needs.  I couldn't disagree more with you about its
"impact".  

> >  Too bad there aren't versions of F-PC for the Mac, Amiga, etc.
> 	FPC is Tom Zimmer's quite brilliant idea of making a Forth
> for the PC/DOS world that runs nearly as well as JForth on the Amiga.
I need to see JForth, obviously.  It must be amazing.  Since you've obviously
seen both I'll defer to your judgement and your suggestion in a previous
posting that it should be JForth which is elevated to a "standard".  I
understand that JForth is a commercial Forth--good for the vendors.  If I
had an Amiga I'd probably be happy to pay for it.  I wonder how the authors
would feel about having the "look and feel" of their implementation cloned
on other architectures?  Wouldn't it be fun if a handful of the most
creative, generous, and prolific Forth implementors (maybe Tom Zimmer,
Roedy Green, Mitch Bradley and the authors of JForth) formed their own
standards committee and introduced a new generation of Fig Forth?
 
> >I'm amazed that the Forth community apparently has such a poor
> >grasp of human psychology as it pertains to motivation and public
> >relations.
> 	Forth programmers, as a class, couldn't sell shaved ice in
> hell. We are working on it. Image will change.
By "we", do you mean Fig?  Is there an ORGANIZED, intelligent effort underway
to change Forth's image?  I (obviously) like campaigns.  Maybe I can
help in some way?

> >go the way of the dinosaurs--just project how far it's fallen from
> >favor over the last decade and project that to the year 2000.  No, you
> >don't need to be a mathematician to figure out where Forth is going
> >to be. 
> 	Forth is in use in more installations than ever before in its
> history. It is becoming the premiere embedded control language, which is where
> it was born.
...
> 90% of the computer bux are spent on embedded systems in America,
> 10% on the type of computer people sit at a desk and use.
  
Surely you know the difference between absolute numbers and market share?  What
you're saying here and in a previous message is that Forth is doing very
well now that it has admitted defeat as a general language and RETREATED to
the niche it came from -- instrument control.  I sure wouldn't count that as a
victory. The number of microcomputers has obviously exploded in the last
decade.  I imagine even the most obscure language can gather statistics
showing that the number of implementations of that language has grown
dramatically.

> 90% of the computer bux are spent on embedded systems in America,
> 10% on the type of computer people sit at a desk and use.
Statistics can be abused to "prove" all sorts of things.  I'm curious to know
how this was calculated but I will accept it, for now, as gospel truth.
So...there are at least ten million PCs out there (probably many more)?
How many embedded control systems?  What MARKET SHARE does Forth have in
embedded control systems?  It's such a great language and it's been around
for almost 20 years and it's gotten a lot of press.  Surely Forth must have
at least 50% -- hopefully more like 90% -- market share?  Are you saying
that there are millions and millions of embedded Forth systems out there or
only a few very expensive ones?  You mentioned earlier that your company
had something like a 20% annual growth over the last five years.  That surely
isn't peanuts but, speaking frankly, it isn't exactly super impressive,
either.  In fact, I have a hunch that if you segregated microcomputer software
companies into three camps -- declining, stable, growing -- you'd find that
20% growth is kind of average for growing companies.  Just a hunch, you 
understand.  But I've spoken to employees of a number of companies which are
advancing in the Unix software market and they are reporting 25% revenue
growth rates as sort of an average.  There are apparently some companies
which are doing much better.

And if Forth DOESN'T have a 50% or better market share in embedded systems
(tell me...does it even have 10%???) what is it going to do when -- five
or ten years from now -- all embedded systems...even toasters...sport at
least a meg of memory?  Can Forth beat out C in such a market when practically
all computer programmers know C (having learned it on the "10%" of the
computer market which Forth has conceded) and almost nobody knows Forth?
-- 
{pacbell!sactoh0! OR ucdavis!csusac!}tree!stever