[ColorForth] Chuck and Browser philosophy

["Kurt B. Kaiser" <kbk@xxxxxxxxx>]

Jeff Fox wrote:
> 
> "Kurt B. Kaiser" wrote:
> 
> > for Colorforth??  Mumble. There's an impedance mismatch here....
> 
> But you missed my whole point.  Trying to match the
> idea of colorForth to that IS a total impedance mismatch.
> Absolutely.  The big mismatch in intellectually.
> ColorForth really has zero to do with any of that.
> ColorForth was and is the perfect match to what
> it was intended for, Chuck's chips.  You put the
> two together it makes sense.

No, no, I understand exactly what you are saying.  What I'm saying is that one
of these superboxes is way overkill for hosting a Colorforth environment.  On
the other hand, I can buy one for the same price you are charging for an F21d
prototype. The market for millions of superboxes has brought their cost down
incredibly. 

In fact, it's fairly difficult to host Colorforth on a modern PC because of the
intricacies of protected mode, memory managers, multilevel caches, multiple
busses, and zillions of hardware configurations requiring special drivers. At
the end of the day, do I really understand the hardware completely once CF is
installed?  I'd much rather be able to afford to do it on something simpler :)

Even a PDA is hundreds of dollars for a color version, and the user interface
stinks.

Although suffering from the same problems, a junk 386 would seem to be an
adequate platform for CF. But I seem to recall that Colorforth currently
requires a Pentium, something about a CPU time stamp? 

I'm running Colorforth on a DEC Celebris GL which is a Pentium Pro 200 with
Matrox graphics. I got three of them for $90, so I have spare parts. CF came
right up on it with no problems.

On the other hand, those $100 8 bit development platforms are underkill, IMO.
Maybe the TI MSP430 development module for $99 that Forth Inc is interfacing
to?

Isn't there a company selling one of Chuck's chips disguised as a non-Forth
product? Could it still run Forth? Is there a cheap way to interface one to a
display and I/O? I looked for it on your chip page but missed it.

> You have a $2-internet-colorForth that can change
> everything internally.
> 
> To pull himself up by his bootstraps Mr. Moore
> is temporarily using a rather personal cludge
> on the horrible Pentium, as painful as that
> is for Chuck, but it shows what he wants
> and were he wants to go.

Yes, he's stuck with the rest of us.  The trick is to spend the minimum of time
to overcome the excessive complexity of the Wintel platform. Seems to me we
spend _most_ of the time dealing with block device I/O and video I/O details,
as opposed to developing Colorforth. It was easier on my Apple II, maybe I'll
dig it out of storage.

And how do you get ColorForth to print out code on your color inkjet? Ye Gods.

> Look closely and you see the idea.  If the
> ideas you see is that this is Chuck's
> personal language implemention for a Pentium
> and that if you apply the idea on top of
> Windows or Linux then you will be doing the
> same thing Chuck is doing but that it is
> obvious that Chuck's approach has all these
> flaws, and problems that make it nearly
> impossible.
> 
> It is nearly impossible and won't work.
> Chuck's approach isn't a personal system.
> Chuck's approach isn't intended for Pentium.
> Chuck's approach can't mate to Windows or Linux
> the whole idea from the beginning is to avoid
> and replace all that.
> 
> So while by Chuck's definition of the concept
> of colorForth vs other Forths you could
> do a personal Pentium colorForth for an
> antithesis OS and call it colorForth IMHO
> if this is your target your idea is almost
> exactly the opposite of Chuck's.  You
> think you are seeing what Chuck is doing
> and that you are doing the same thing.
> You are both using the word colorForth
> for instance.
> 
> But unless you feel that touching the Penitum

I assume you mean "But if you feel...."

> in the first place is terribly distasteful,
> but not as distastful as touching Windows
> or Linux and feel so passionate about that
> that you are willing to risk everything that
> you have done for 40 years on trying to do
> something else then you are seeing ColorForth
> and doing what Chuck is doing.
> 
> If you are using colorForth the way Chuck is
> it is presently just a temporary phase required
> for CAD, to make chips, to complete the picture
> to get to the real colorForth.  If your goal
> is a $2-internet-colorForth computer and

I assume you mean "If your goal isn't...."

> operating system that is the anti-thesis
> of "personal" software, and the anti-thesis
> of things like Linux your not doing what
> Chuck is doing and are missing the message
> almost everything.

There's two reasons for implementing ColorForth. First, Chuck's, to use as a
tool to develop his chips. The rest of us can't do that. Second, to experiment
with new concepts in Forth and language/OS design, especially as related to
"low fat" systems. The trick here is to avoid letting the screwing around with
the host platform dominate our activity. The goal would be to develop
enhancements/applications so that when the great day comes and we can boot up a
device (with, say, a Forth chip, display, keyboard, mass memory, and a way to
get information in/out) we can pick up our knowledge and applications and
install them on the new hardware.

Remember that Woz did the Apple on his living room floor. Designed the hardware
and Woz machine floppy controller, implemented ROM Basic and the monitor. Then
he sold kits for awhile.

Netpliance (NPLI) got $133 million in their IPO for a network box and blew half
of it before they gave up on that business. They _scrapped_ a warehouse full of
hardware, each box essentially a PC. (Curiously, they were originally
incorporated as Shbang!, any relation to Shboom?)

"Only" a couple of hundred thousand bucks is needed to push the chip project
along. 

"The best is the enemy of the good."

> 
> Until you picture how every detail of the
> $2-internet-colorForth computer and OS
> will work, it is hard to see that it is
> the anti-thesis of personal software
> or things like Unix.  If that isn't your
> image IMHO you may have only a very
> surface understanding of the concepts
> or possibly have most of it backwards.
> Not "you" personally, anyone.
> 
> > I wouldn't toss TC/IP,
> 
> Not even to replace it with something significantly
> better if doing so opened the door to get
> 100,000% improvement in some other things?
> 
> My whole point was most people will not even
> consider ideas like that so they just will
> never be able to get more than 10% of Chuck's
> ideas.  His ideas are so compact and succinct
> that if you remove about 1% most of the other
> 99% becomes broken.  There is no fat to
> remove, you are removing vital parts.
> 
> Chuck's approach is about removing the fat
> from things that are mostly fat and making
> them with no fat and only essential parts.
> So people who remove essential parts from
> what Chuck made, all they have is a broken
> toy, and the parts can't fit into anything
> else.

What specific essential parts being removed are you referring to?

> > you don't want to build your own Internet :) But most
> > everything else could go, in principle.
> 
> You may not want to.  Chuck does.  I do.  5.9 billion
> other people would like to see it done.  I suppose the
> 100 million who use the present would be split
> between those wanting to replace it with a better
> version and those who would resist change.  And
> there would be almost none who would ever consider
> doing it themselves.

There is no problem carrying ColorForth(OS) proprietary messages via TCP/IP
datagrams. If you don't use the Internet and its IP routers, what are you
planning to use for connectivity? Rather than start from zero, you want to
insinuate your technology into the existing stuff. After all, TCP/IP was
designed over 30 years ago in a simpler era. It is not excessively complicated
for what it does. Show me a better design.

You will never prosper if you believe that the rest of the world never had any
good ideas or made any progress. They did. Steal the best and dump the rest.

> For Chuck I say why not?  If your $2-internet-colorForth
> can generate a better internet give it a try.
> 
> > Did you see the communal computers being used in India?
> > About the size of a large PDA. A town owns one, and each
> > user has a plug in card containing his
> > configuration and data. Different paradigms!
> 
> Yeah, closer to Chuck's idea.  Chuck just wants the
> PDA to have a $2-internet-colorforth to get orders
> of magnitude more performance and lower cost so
> everything can be more efficient.  What you describe
> sounds a little like other technologies 100 years
> ago.

I appreciate your comments on events at iTV in your previous posts.  Do you
have any more "lessons learned" to relate?

Regards, KBK
-- 
   K u r t    B.   K a i s e r
   k b k @@ s h o r e .. n e t
------------------------------
------------------------

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