[ColorForth] Cross purposes.....

[Jeff Fox <fox@xxxxxxxxxxxxxxxxxxx>]

David Govoni wrote:
> My own personal gut feeling is that the 25X or 50X 
> will (or would) be more successful as a programmable 
> and powerful coprocessor for existing systems
> than as a stand alone device. 

I was intended for that, and commissioned for
that.  The idea was to use C on support Windows,
Linux and BSD.  Drop in product.

But it was apparent to me that Chuck also
invested considerable effort to make the
outside of the chip, almost the entire
I/O achitecture, programmable. Almost
like bying a BIG FPGA with a hard CPU
core or cores to be more precise and an 
undefined interface to the outside world.

I saw a chip like that the other day.
I thought.  Geez, can you imagine them
spending the billions that they did to
make it only to have the entire world
somehow miss the whole idea behind the
product and all be happy to get 1
version with the programmables all
programmed for everyone in the world
in one form.  Why build programmable
hardware and then program it only one
way?  That is what dedicated circuits
are for.  If you use the FPGA as one part
it is a very expenive way to get one part
and a very inefficient part constructed
with gates that carry overhead to be
programmable for no reason.

Anyway I have just been trying to provide
some ideas.  Like Chuck, I try to toss out
a lot of them.  A lower percentage of mine
are good than Chuck's. I am sure a lot of
the ideas are poor.  So are some of Chuck's
but he so many more than most people that
he also has about 100 times as many good
ideas as they do too.

In fact the colorForth list never really
interest me.  I was worried about where
it might go.

I can understand that chuck's colorForth
is interesting.  I am intersted enough to
also want to know more.  I am interested
in others people's too.  I encourage
experimentation too, it is an idea in
its infancy.  Who knows where it will go.

I would love to see lots of people discover
ideas that never occured to me.  But Chuck
says USING is most important and he USES
it for OKAD but we can't...  But he also
uses it to do a simulator and cross compiler
for his chip.   He is bootstrapping.  The
simulator and cross compiler are step 2
to the real vision of what colorForth
and where it is goning.

I am of the opinion that the best colorForth
to study now is the target compiler in it
and the simulator and using the target
compiler and simulator.  Using colorForth
more to real colorForth.  How can one
really understand colorForth from only
the Pentium version?  It is not only full
of ugly stuff required on Pentium, but
understanding Pentium is 1000 times
more difficult than understanding the
target chips which when understood
cause the real ideas behind colorForth
to make real sense.

I worry that we will just get too many
new branches before the several step
process is done and everyone will be
gone off on their own ideas of colorForths
all off in their own directions.  We will
have something that looks like that ANS
branch, but may be worse, because not only
will it splinter but there will be little
if any sharing because people were happy
with that idea from the start.

Some people are delighted to scape a little
of the shiny surface of Forth off, something
like Forth factoring, and add it on top of
crap.  They get shiny crap.  Ideas from
Forth can be lifted and used in other languages.
It isn't a bad thing.  Shiny crap is better
than dull crap I guess.

I think a shining jewel is even better however
so I say don't split the diamond and put the
dust on the jewelry, use the diamond since it
is far more beautiful.  But diamond dust is
better than nothing or glass.

So I worry that it sounds like everyone sees
Chuck as loner and is happy with their own
lowner versions.  Chuck complains that people
do it, because he wants them to do more,
and encourages them to do it because he thinks
that they not be capable or may not want to.
And that is fine.  Do your won thing. I like DIY.

But what I would like you to do, yourself, is
do what Chuck would if he had the time and try
you best to support him by writing code that
will help.  Writing your own versions or
playing with his and reading his site and this
list are good places to start.  

After a while consider trying the deep end.
Try writing software, in colorForth the
target chip.  It can't hold much.  You
don't have to write megabytes.  A
few dozen words is plenty to simulate
and see what it does on the target chip
that is the other half of what Chuck's vision
is all about.

I might even like a colorForth style editor
for ANS Forth, or a colorForth style editor
in Aha.  I am behind most of you folks on
this whole colorForth subject.

But I think I graps the theories and know about
some of the implemenation techniques that Chuck
has tried so far and that he is currently using
and could change tomorrow.  But I personally
would rather focus on something other than
Pentium, F21 and know that other people don't
have those.  But for years using a simluator
for me was fine and almost as much fun.  And
Chuck has now published both the cross compiler
for c18 and simulator for 25x.  That is where
my first interests would be, but that's me.

After seeing how the ideas evolved half in
hardware and half in software I really do
feel that this the only way to really
understand them.  Most people want to
concentrate on the software or hardware,
and with enough layers to separate them
and enough people isolated from one another
on different levels people do specialize
on only one level.  The hardware folks
want to isolated from software and visa-versa
but the fact they are both sides of the
same coin here makes specializing too
much and simply to not even know about the
other side of the coin just seems so silly
to me.  

Don't you want to understand what you are doing?   
Would you set out to study assembler by refusing to learn anything
about the hardware like the instruction set
and architecture?

Chuck put in an effort to do that so that
the chip would not be limited to only
be used on boards that drop into PCs
or go into racks and get controlled
and fed data by PCs.  You know that
Chuck was have little interest in being
involved directly with that work after
delivering the chip and some help.
After that all the work is Windows and
C and Chuck isn't interested in doing that
he will be doing something else.  Probably
NOT using the chip for any of the billion
other things it could do.

Because of its programmable hardware
interface and the nature of that I keep
coming up with new ideas.  But then I
realized that I was the only ONE who
thought about the chip that way!

It is SO frustrating when you see a
brilliant idea, think it is the greatest
thing you have seen in 14 months on
the project and get really excited
that perhaps that is what is needed.

Sure it has limits, the limits of the
cleverness of the programmers who if they
wanted to could turn it into a million
new incredible new machines that no one
has ever imaged, or millions of drop in
replacments that are cheaper or better
for devices that already exist.

I think this is the best development yet.
Chuck does one piece of hardware and 1000
other interested people can each think
up 10 ideas and we ahve 10,000 new paths
to make it a success. Some of them will
have to succed.

Then I realize, not a single person seems
to get it.  Everyone only sees it as with
an SRAM bus and as a drop in part for
windows.  

I realized that once again the incredible power
of Chuck's ideas is totally wasted no one
gets that their are a million different ways
that the chip can be configured and that
Chuck was probably expecting what he calls
"collaberative developemnt" from these
people as he did from me.

So because they are limited by the fact that
this is something new and they are like to only
be able to see it framed by what they already
know and all those dreadfull illusions that
everyone is burdened with and that limit them.
I thought well once again I suspect that
Chuck overestimated others.  The value of
his chip really IS 10,000 lower than his
expectations simply because he didn't count
on everyone's illusions preventing them from
seeing that this is 10,000 new chips, not one.
So his 1,000 collaberative helpers don't exist.
By being comfortable with the limitations
and are excited about 1 new chip, that's
what we will have.  Their illusion will become
real because they believe it, 99.9% of Chuck's
efforts will be wasted by this.  Maybe if
I jump in right away and wake them up!

I am really worried about them.  Instead of
turning Chuck's design into the 10,000
chips he imagined, and providing the software
to do that for him as he hoped these people
are far less ambitiout than Chuck.  They
are happy to get 1 chip, in fact, this is
the colorForth list they don't care about
Chuck's chips or MachineForth.  They only
get about 1% of Chuck's ideas in .1% of
the niches he intended.  God, it must be
lonely for Chuck way out there.  Can I
help in any way.

Like people who would be happy to get 1
chip and don't want to think about they
might help Chuck by just adding their software
exptise, and with MachineForth and colorForth
methods it will be easy.  These people are
happy to get one colorForth. Either a copy
of Chuck's or their own.  

In a sense I feel just like vendors who hate
DIY.  Here are all these people that Chuck
hoped would get colorForth.  To me getting
it means getting it, like Chuck.  Seeing
how the ideas have 100% of their value
the way they fit together.  Instead of
accepting that they are going to pull out
5% and mix with the conventional world
and do what everyone else does....

Like being happy with a chip they will be
happy with 5% of colorForth.  They are
coming in with the attitude that it is
a personal language and to be like Chuck
they should go off and "experiment" after
all Chuck says that that is good.

He also says that doing some with Forth
is GREAT.  He wishes that people would use
it more than anyone.  Once again, I get tired
of the fact that the illusion of most people
is that a few people copying Chuck by being
hermits who get more and more out of sync
with each other because that's what they
are suppose to want.  

Is a score of 1 on a test good?  It is better
than 0.  Congratuations to anyone who has a
1 or higher.  I don't think that people would
be here without a 1 or higher.

I feel Chuck would want people to aspire to
greatness.  I know Chuck does.  If they want to
copy Chuck try that.  Try for 100.  We claim
it would be easy for children.  Why are professional
adults satisfied to aim for 1% ?

But, I don't want to drive anyone away for not
being more ambitious.  I am just trying to
encourage everyone to strive for execellence.
If you get 100% then you have lots of stuff
to play with.  If you decide to only pick
1% to use on your project great now you
have 100 more to pick from.

If everyone accepts that 1% is all that is
worth going for that is all there will be.
There will be 1 chip with software from Chuck.
There will be 1 personal colorForth for Chuck.
There will be 1 personal colorForth from xxx
and 1 from xyz and one from abc and everyone
will by happy watching the population drop
to zero in a few years as they tire of
trying do do much with only their personal
language.

Once again, 1 is better than 0.  It is almost
infinitely better than 0.  Please don't let
me drive anyone away because I encourage you
to try for more.  It's mostly just a matter
of dropping illusions and it doesn't hurt.




Of course that is precisely what it was designed for.

I have assume that when talked about 
The ability to throw 10 or 20 of them into a
> system as general or specific purpose processors for $40 makes my mind go
> through loops. Incredibly powerful and cheap next generation video game
> console anyone? If selling the chip to others for their own use leads to
> Charles being able to create his own system based on it, all the more power
> to him. I'd love to have the time and oppurtunity to program the 25X and
> put it through some paces.
> 
> But I'm on this list because of my interest in ColorForth and Flux and the
> ideas I'm learning from them.
> 
> I'm here to watch Charles, Jeff, Sean and all the others try things and
> share what they've learned. And as I do more work in these systems I want
> to share my ideas and discoveries.
> 
> In a place where we're not at each others throats. Even if we don't agree.
> 
> So, after my minor diatribe I'm still going to post my 3 parter on what
> I've learned, and what I might try to do. I'm looking for comments and
> criticisms. Constructive criticism please, but feel free to let me know
> what you think. Regardless.
> 
> Thanks in advance,
> David Govoni
> 
> ------------------------
> 
> To Unsubscribe from this list, send mail to Mdaemon@xxxxxxxxxxxxxxxxxx with:
> unsubscribe ColorForth
> as the first and only line within the message body
> Problems   -   List-Admin@xxxxxxxxxxxxxxxxxx
> Main ColorForth site   -   http://www.colorforth.com
------------------------

To Unsubscribe from this list, send mail to Mdaemon@xxxxxxxxxxxxxxxxxx with:
unsubscribe ColorForth
as the first and only line within the message body
Problems   -   List-Admin@xxxxxxxxxxxxxxxxxx
Main ColorForth site   -   http://www.colorforth.com