Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.forth > #1049 > unrolled thread

Forth on the Maximite

Back to article view | Back to comp.lang.forth

Contents

  Forth on the Maximite Andrew Reid <andrewreid69@gmail.com> - 2011-04-07 12:47 -0700
    Re: Forth on the Maximite Paul Rubin <no.email@nospam.invalid> - 2011-04-07 13:21 -0700
      Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-10 20:32 -0700
        Re: Forth on the Maximite rickman <gnuarm@gmail.com> - 2011-04-11 08:13 -0700
          Re: Forth on the Maximite anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2011-04-11 15:37 +0000
            Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-11 14:00 -0700
              Re: Forth on the Maximite rickman <gnuarm@gmail.com> - 2011-04-11 15:42 -0700
                Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-11 21:35 -0700
              Re: Forth on the Maximite anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2011-04-12 11:20 +0000
                Re: Forth on the Maximite Bernd Paysan <bernd.paysan@gmx.de> - 2011-04-12 18:07 +0200
          Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-11 11:48 -0700
        Re: Forth on the Maximite Paul Rubin <no.email@nospam.invalid> - 2011-04-12 00:17 -0700
          Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-12 10:19 -0700
            Re: Forth on the Maximite Paul Rubin <no.email@nospam.invalid> - 2011-04-16 12:50 -0700
              Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-16 15:05 -0700
                Re: Forth on the Maximite rickman <gnuarm@gmail.com> - 2011-04-16 15:34 -0700
                  Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-16 16:37 -0700
                    Re: Forth on the Maximite rickman <gnuarm@gmail.com> - 2011-04-16 16:55 -0700
                      Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-16 17:41 -0700
                        Re: Forth on the Maximite rickman <gnuarm@gmail.com> - 2011-04-17 14:38 -0700
                          Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-17 21:08 -0700
                    Re: Forth on the Maximite kenney@cix.compulink.co.uk - 2011-04-17 14:14 -0500
                      Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-17 14:08 -0700
                        Re: Forth on the Maximite kenney@cix.compulink.co.uk - 2011-04-18 04:00 -0500
        Re: Forth on the Maximite Chris Baird <abuse@brushtail.apana.org.au> - 2011-04-16 20:12 +1000
          Re: Forth on the Maximite rickman <gnuarm@gmail.com> - 2011-04-16 15:36 -0700
            Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-16 17:06 -0700
              Re: Forth on the Maximite rickman <gnuarm@gmail.com> - 2011-04-17 14:15 -0700
                Re: Forth on the Maximite Albert van der Horst <albert@spenarnc.xs4all.nl> - 2011-04-18 18:12 +0000
                  Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-18 11:15 -0700
                  Re: Forth on the Maximite rickman <gnuarm@gmail.com> - 2011-04-18 21:44 -0700
                    Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-19 00:23 -0700
                      Re: Forth on the Maximite rickman <gnuarm@gmail.com> - 2011-04-19 07:23 -0700
                        Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-19 10:24 -0700
                          Re: Forth on the Maximite rickman <gnuarm@gmail.com> - 2011-04-19 12:22 -0700
                            Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-19 16:14 -0700
                              Re: Forth on the Maximite rickman <gnuarm@gmail.com> - 2011-04-20 00:08 -0700
                                Re: Forth on the Maximite stephenXXX@mpeforth.com (Stephen Pelc) - 2011-04-20 10:46 +0000
                                  Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-20 14:55 -0700
                                Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-21 20:24 -0700
                                  Re: Forth on the Maximite BruceMcF <agila61@netscape.net> - 2011-04-21 21:02 -0700
                                    Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-21 23:09 -0700
                                      Re: Forth on the Maximite "David N. Williams" <williams@umich.edu> - 2011-04-22 08:42 -0400
                                      Re: Forth on the Maximite BruceMcF <agila61@netscape.net> - 2011-04-22 07:06 -0700
                                        Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-22 12:18 -0700
                              Re: Forth on the Maximite Andrew Haley <andrew29@littlepinkcloud.invalid> - 2011-04-20 04:07 -0500
                    Re: Forth on the Maximite Albert van der Horst <albert@spenarnc.xs4all.nl> - 2011-04-19 19:05 +0000
                  Re: Forth on the Maximite Micke <oh2aun@gmail.com> - 2011-04-19 12:00 -0700
          Re: Forth on the Maximite Bluebee <visualforth@rocketmail.com> - 2011-04-16 15:52 -0700
            Re: Forth on the Maximite Chris Baird <abuse@brushtail.apana.org.au> - 2011-04-17 23:01 +1000
        Re: Forth on the Maximite Hugh Aguilar <hughaguilar96@yahoo.com> - 2011-04-21 19:52 -0700
      Re: Forth on the Maximite Hugh Aguilar <hughaguilar96@yahoo.com> - 2011-04-10 21:58 -0700

Page 1 of 3  [1] 2 3  Next page →

#1049 — Forth on the Maximite

There is a new exciting microprocessor on the market, the PIC32, based
on the famous MIPS architecture. In Australia I got a little box
called "MAXIMITE" with this micro, and it would be great to have Forth
on it. Is anyone out there to help me to get a MIPS Forth for the
PIC32?

It has GWBasic so a new generation may be interested in interpreter &
Compiler facilities to aid real time Dataloging and control
applications.  Costs approx $A79 in Kit form.  Forth would be
perfect!!

[toc] | [next] | [standalone]

#1051

Andrew Reid <andrewreid69@gmail.com> writes:

> There is a new exciting microprocessor on the market, the PIC32, based
> on the famous MIPS architecture.

Is something like that really exciting in this day and age, given the
pervasiveness of ARM stuff?

> In Australia I got a little box called "MAXIMITE" with this micro, and
> it would be great to have Forth on it. Is anyone out there to help me
> to get a MIPS Forth for the PIC32?

I'd expect there's a GCC toolchain and you can port Gforth.

[toc] | [prev] | [next] | [standalone]

#1140

On 7 Apr., 16:21, Paul Rubin <no.em...@nospam.invalid> wrote:
> Andrew Reid <andrewrei...@gmail.com> writes:
> > There is a new exciting microprocessor on the market, the PIC32, based
> > on the famous MIPS architecture.
>
> Is something like that really exciting in this day and age, given the
> pervasiveness of ARM stuff?
>

Paul, I was wondering, too, because I like the ARM micro and I
remember very well that once an ARM manager was interviewed and he
said (with my own words here): "The ARM architecture will prevail and
ARM microprocessors always will be cheaper and need less energy than
others because of its low gate count!"

But now the situation seems to change. The MIPS processor like the
ARM, shares many of the qualities found with a RISC processor – high
performance with low power consumption, low gate count and small die
size. And believe it or not, MIPS cores can be found in more than two
thirds of newer Cisco routers, cable modems and ADSL modems,
smartcards, laser printer engines, VoIP applications, set-top boxes,
DVD recorders, robots, handheld computers, Sony PlayStation 2 and Sony
PlayStation Portable. MIPS seems to be on the rise.

That wouldn't be a reason for me to chose a MIPS microprocessor, but
microchip has a unique standing in encouraging young adults to use
microprocessors and supporting them.

Still sceptical and still believing that the ARM will be a better
choice, I looked for comparable ARM microprocessors from different
vendors. And to my surprise, there is no ARM microprocessor available
at that low price with that high amount of memory (512k program, 128k
RAM), UART/SPI/I2C, RTCC, 16 AD channels, 5 Timers/PWM, USB/CAN/
ETHERNET, 8 DMA's, and only 64 pins. Microchip licensed the MIPS
architecture, a really powerful machine, and because of Microchips
high market penetration this will be the right micro for starting
Forth, unbeatable in price.

> > In Australia I got a little box called "MAXIMITE" with this micro, and
> > it would be great to have Forth on it. Is anyone out there to help me
> > to get a MIPS Forth for the PIC32?
>
> I'd expect there's a GCC toolchain and you can port Gforth.

Now I did my second research looking for GCC toolchain and Gforth.
The GCC toolchain told me that porting Gforth would be a daunting task
for me.

But reading the Gforth goals, I found these words: "It [Gforth] should
become standard, i.e. widely accepted and used. This goal is the most
difficult one." In my opinion the PIC32 is the window of opportunity
to reach this goal.

I installed Gforth on my laptop, it works really well.
And then the next surprise: the source code  includes a MIPS
assembler!!!

MIPS IV, which is the PIC32 CPU, is the fourth version of the
architecture. It is a superset of MIPS III and is compatible with all
existing versions of MIPS.
(Source:  http://en.wikipedia.org/wiki/MIPS_architecture).

It would be great if somebody will take the job to make Gforth run on
the PIC32.
As I wrote, a window of opportunity. And most of the work seems to be
done.

[toc] | [prev] | [next] | [standalone]

#1152

On Apr 10, 11:32 pm, Bluebee <visualfo...@rocketmail.com> wrote:
> On 7 Apr., 16:21, Paul Rubin <no.em...@nospam.invalid> wrote:
>
> > Andrew Reid <andrewrei...@gmail.com> writes:
> > > There is a new exciting microprocessor on the market, the PIC32, based
> > > on the famous MIPS architecture.
>
> > Is something like that really exciting in this day and age, given the
> > pervasiveness of ARM stuff?
>
> Paul, I was wondering, too, because I like the ARM micro and I
> remember very well that once an ARM manager was interviewed and he
> said (with my own words here): "The ARM architecture will prevail and
> ARM microprocessors always will be cheaper and need less energy than
> others because of its low gate count!"
>
> But now the situation seems to change. The MIPS processor like the
> ARM, shares many of the qualities found with a RISC processor – high
> performance with low power consumption, low gate count and small die
> size. And believe it or not, MIPS cores can be found in more than two
> thirds of newer Cisco routers, cable modems and ADSL modems,
> smartcards, laser printer engines, VoIP applications, set-top boxes,
> DVD recorders, robots, handheld computers, Sony PlayStation 2 and Sony
> PlayStation Portable. MIPS seems to be on the rise.

Yes, Cisco picked a processor to use in their ASICs and once they made
that decision, they are working with it.  Does that mean it was the
right decision?  Britain and a number of other countries picked the
left side of the road to drive on and are now living with that
decision.  Does that mean it was the right decision?  It means that
changing is a large investment, in fact more than the original
investment in picking one processor over the others since there would
be a huge cost related to the confusion factor were they to change.


> That wouldn't be a reason for me to chose a MIPS microprocessor, but
> microchip has a unique standing in encouraging young adults to use
> microprocessors and supporting them.

No it is not a reason to change, in fact, I find it to be fairly
irrelevant.  As to Microchip's "standing", I lament it often.  The PIC
architecture is not a good one so that newbies would be much better
off learning a processor that has a more regular instruction set.  I
suppose, again, it is more a matter of inertia and resistance to
change.  Many newbies learn from other newbies so that they often
follow the same learning path even if it is harder than necessary.


> Still sceptical and still believing that the ARM will be a better
> choice, I looked for comparable ARM microprocessors from different
> vendors. And to my surprise, there is no ARM microprocessor available
> at that low price with that high amount of memory (512k program, 128k
> RAM), UART/SPI/I2C, RTCC, 16 AD channels, 5 Timers/PWM, USB/CAN/
> ETHERNET, 8 DMA's, and only 64 pins. Microchip licensed the MIPS
> architecture, a really powerful machine, and because of Microchips
> high market penetration this will be the right micro for starting
> Forth, unbeatable in price.

That is an amazingly narrow basis for picking one processor family
over another.  You picked one obscure, very specific configuration and
compared the price point.  That tells you that for this one
combination of features the ARM processor is not the best choice...
so?  What about the other 2**256 other combinations of features?  If
you leave out the Ethernet I don't think the PIC32 is the best choice
anymore.  If you require more than 12 bits of resolution I believe
that leaves the PIC32 out entirely!


> > > In Australia I got a little box called "MAXIMITE" with this micro, and
> > > it would be great to have Forth on it. Is anyone out there to help me
> > > to get a MIPS Forth for the PIC32?
>
> > I'd expect there's a GCC toolchain and you can port Gforth.
>
> Now I did my second research looking for GCC toolchain and Gforth.
> The GCC toolchain told me that porting Gforth would be a daunting task
> for me.
>
> But reading the Gforth goals, I found these words: "It [Gforth] should
> become standard, i.e. widely accepted and used. This goal is the most
> difficult one." In my opinion the PIC32 is the window of opportunity
> to reach this goal.

I think you will find porting a Forth to the ARM processor will result
in it be use much, much more!


> I installed Gforth on my laptop, it works really well.
> And then the next surprise: the source code  includes a MIPS
> assembler!!!
>
> MIPS IV, which is the PIC32 CPU, is the fourth version of the
> architecture. It is a superset of MIPS III and is compatible with all
> existing versions of MIPS.
> (Source:  http://en.wikipedia.org/wiki/MIPS_architecture).
>
> It would be great if somebody will take the job to make Gforth run on
> the PIC32.
> As I wrote, a window of opportunity. And most of the work seems to be
> done.

Please don't misunderstand me.  If someone wants to port gforth to the
MIPS processor that is a great thing.  I am just pointing out that I
think the reasons you state for picking the MIPS architecture are not
completely accurate.  I personally would much prefer to see forths
ported to the ARM architecture because of its pervasiveness.  I think
in the grand scheme of things the differences between the processors
themselves won't amount to a hill of beans.  But ARM MCUs are being
sold by how many dozen vendors now?  The MIPS is still just what, one,
two, maybe three?  If you want flexibility and selection in MCUs, the
ARM is the way to go.

One other point.  Porting any PC oriented forth to an MCU is nothing
like porting between operating systems.  What is needed I think is a
Forth cross-compiler that targets a given MCU and I think that is a
very different animal requiring a lot of work... if it wasn't I think
we would already see at least as many cross-compilers as there are
forths.  There is one for the ARM, http://pygmy.utoh.org/riscy/

Rick

[toc] | [prev] | [next] | [standalone]

#1154

rickman <gnuarm@gmail.com> writes:
>Please don't misunderstand me.  If someone wants to port gforth to the
>MIPS processor that is a great thing.

That work has already been done, and Gforth also runs on many other
architectures, ARM among them.  My guess is that the PIC32 and similar
embedded targets would still require some porting work, not because of
the architecture, but because of the software environment.

- anton
-- 
M. Anton Ertl  http://www.complang.tuwien.ac.at/anton/home.html
comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
     New standard: http://www.forth200x.org/forth200x.html
   EuroForth 2010: http://www.euroforth.org/ef10/

[toc] | [prev] | [next] | [standalone]

#1161

On 11 Apr., 11:37, an...@mips.complang.tuwien.ac.at (Anton Ertl)
wrote:
> ... Gforth also runs on many other architectures, ARM among them.
> ... My guess is that the PIC32 and similar
> embedded targets would still require some porting work, not because of
> the architecture, but because of the software environment.
>
> - anton
> --
> M. Anton Ertl  http://www.complang.tuwien.ac.at/anton/home.html
> comp.lang.forth FAQs:http://www.complang.tuwien.ac.at/forth/faq/toc.html
>      New standard:http://www.forth200x.org/forth200x.html
>    EuroForth 2010:http://www.euroforth.org/ef10/

Since I started programming microprocessors in 1975, I always worked
on the target microprocessor system itself, using the microprocessors
monitor (the first one was the KIM) and then a (later virtual)
terminal. Since I started with Forth nine years later, I never needed
breakpoints or single step debugging any more. That was for me one of
the beauties of Forth.

What is the problem with the software environment?
What steps have to be done to get Gforth running on a MIPS micro?

[toc] | [prev] | [next] | [standalone]

#1165

On Apr 11, 5:00 pm, Bluebee <visualfo...@rocketmail.com> wrote:
> On 11 Apr., 11:37, an...@mips.complang.tuwien.ac.at (Anton Ertl)
> wrote:
>
> > ... Gforth also runs on many other architectures, ARM among them.
> > ... My guess is that the PIC32 and similar
> > embedded targets would still require some porting work, not because of
> > the architecture, but because of the software environment.
>
> > - anton
> > --
> > M. Anton Ertl  http://www.complang.tuwien.ac.at/anton/home.html
> > comp.lang.forth FAQs:http://www.complang.tuwien.ac.at/forth/faq/toc.html
> >      New standard:http://www.forth200x.org/forth200x.html
> >    EuroForth 2010:http://www.euroforth.org/ef10/
>
> Since I started programming microprocessors in 1975, I always worked
> on the target microprocessor system itself, using the microprocessors
> monitor (the first one was the KIM) and then a (later virtual)
> terminal. Since I started with Forth nine years later, I never needed
> breakpoints or single step debugging any more. That was for me one of
> the beauties of Forth.
>
> What is the problem with the software environment?
> What steps have to be done to get Gforth running on a MIPS micro?

Do you run Forth on micros with less than a couple K of RAM or just a
few K of program memory?  I suppose I don't expect MIPS chips to have
such low amounts of memory, but I guess I just expect a system to
support a good range.  Maybe that's not so important for this
approach. It certainly would make it a much more simple effort.

Rick

[toc] | [prev] | [next] | [standalone]

#1169

On Apr 11, 6:42 pm, rickman <gnu...@gmail.com> wrote:
> On Apr 11, 5:00 pm, Bluebee <visualfo...@rocketmail.com> wrote:
>
> > On 11 Apr., 11:37, an...@mips.complang.tuwien.ac.at (Anton Ertl)
> > wrote:
>
> > > ... Gforth also runs on many other architectures, ARM among them.
> > > ... My guess is that the PIC32 and similar
> > > embedded targets would still require some porting work, not because of
> > > the architecture, but because of the software environment.
>
> > > - anton
> > > --
> > > M. Anton Ertl  http://www.complang.tuwien.ac.at/anton/home.html
> > > comp.lang.forth FAQs:http://www.complang.tuwien.ac.at/forth/faq/toc.html
> > >      New standard:http://www.forth200x.org/forth200x.html
> > >    EuroForth 2010:http://www.euroforth.org/ef10/
>
> > Since I started programming microprocessors in 1975, I always worked
> > on the target microprocessor system itself, using the microprocessors
> > monitor (the first one was the KIM) and then a (later virtual)
> > terminal. Since I started with Forth nine years later, I never needed
> > breakpoints or single step debugging any more. That was for me one of
> > the beauties of Forth.
>
> > What is the problem with the software environment?
> > What steps have to be done to get Gforth running on a MIPS micro?
>
> Do you run Forth on micros with less than a couple K of RAM or just a
> few K of program memory?  I suppose I don't expect MIPS chips to have
> such low amounts of memory, but I guess I just expect a system to
> support a good range.  Maybe that's not so important for this
> approach. It certainly would make it a much more simple effort.
>
> Rick

I started with RSC-Forth (8kByte Kernel). There was enough space left
over for application programs using an 32kBye EPROM, and 32kByte RAM
was enough, too. Later I ported RSC-Forth to Mitsubishi/Renesas M38049
with 2kByte of RAM, but that worked too. When my application needed
extension, it didn't fit into the M38049 FLASH anymore, so I decided
to write a FORTH extension to compile the additional application into
a little serial EEPROM, and run from this EEPROM. Worked really well.

The PIC32 MIPS micro has 512kBytes of program memory and 128kBytes of
RAM. I guess that's enough for sophisticated microprocessor
applications.
DB

[toc] | [prev] | [next] | [standalone]

#1176

Bluebee <visualforth@rocketmail.com> writes:
>On 11 Apr., 11:37, an...@mips.complang.tuwien.ac.at (Anton Ertl)
>wrote:
>> ... My guess is that the PIC32 and similar
>> embedded targets would still require some porting work, not because of
>> the architecture, but because of the software environment.
...
>What is the problem with the software environment?

Gforth uses a number of functions for setting up the memory, for I/O,
and for dealing with exceptional conditions.  It can make do without
some of them, but if any of the others are missing, it requires
changes in Gforth; nothing that can't be done, but it's work.

Also, if you are cross-compiling, some of the auto-configuration stuff
gdoes not work and you have to use manual configuration for these
things.

- anton
-- 
M. Anton Ertl  http://www.complang.tuwien.ac.at/anton/home.html
comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
     New standard: http://www.forth200x.org/forth200x.html
   EuroForth 2010: http://www.euroforth.org/ef10/

[toc] | [prev] | [next] | [standalone]

#1181

Anton Ertl wrote:
> Gforth uses a number of functions for setting up the memory, for I/O,
> and for dealing with exceptional conditions.  It can make do without
> some of them, but if any of the others are missing, it requires
> changes in Gforth; nothing that can't be done, but it's work.
> 
> Also, if you are cross-compiling, some of the auto-configuration stuff
> gdoes not work and you have to use manual configuration for these
> things.

There's an example how to get Gforth working on an embedded processor 
using cross compiling.  It's for ARM and the beagle board, you can find 
it under gforth/arch/arm/beagle.  The cross compilation auto-config 
stuff is found in config.sh (assign the variables), and there are some C 
programs for serial line interfaces and such.  Of course, all that needs 
to be changed for any embedded MIPS architecture, but the infrastructure 
is there.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/

[toc] | [prev] | [next] | [standalone]

#1158

On 11 Apr., 11:13, rickman <gnu...@gmail.com> wrote:
> On Apr 10, 11:32 pm, Bluebee <visualfo...@rocketmail.com> wrote:
>
> > On 7 Apr., 16:21, Paul Rubin <no.em...@nospam.invalid> wrote:
>
> > > Andrew Reid <andrewrei...@gmail.com> writes:
> > > > There is a new exciting microprocessor on the market, the PIC32, based
> > > > on the famous MIPS architecture.
>
>
> No it is not a reason to change, in fact, I find it to be fairly
> irrelevant.  As to Microchip's "standing", I lament it often.  The PIC
> architecture is not a good one so that newbies would be much better
> off learning a processor that has a more regular instruction set.

Reading that Andrew discovered the MAXIMITE (a hobbyist project, sold
as a kit), my first thought was that I never would use a PIC micro,
even if it had a Forth (and there is one available, IIRC), but Andrew
tried to convince me that the MAXIMITE would be worth it. Of course I
looked for arguments against the PIC32 - because it is named PIC - and
to my surprise, it is not that PIC which I hate. It is a MIPS micro.
But still for me the ARM was the better architecture - I had a project
with MPE's Tiny ARM stamp, and it was great, really easy to use -
therefore I was sure that there will be an ARM micro somewhere with
this amount of memory and all this I/O and the low pin count at a
reasonable price, to convince Andrew that such a project would be
better off using an ARM micro. But to my disappointment there was
none. An ARM micro with this amount of memory has a much higher pin
count, and with the same pin count there was half the amount of
memory, nearly double in price. Getting this result my idea was:
answering is worth it.

> suppose, again, it is more a matter of inertia and resistance to
> change.  Many newbies learn from other newbies so that they often
> follow the same learning path even if it is harder than necessary.

I always was wondering why newbies are doing this instead of looking
for better stuff, but here this is just the point! So many people stay
with PIC (sorry, I do not have the real numbers, I only see the media
promoting the PIC again and again, I guess Microchip has an aggressive
PR crew).

Making a Forth available for the PIC32 will use this inertia and
resistance to change.
And suddenly these newbies discover that using Forth is a unique
adventure, a great pleasure, and makes programming so much more
powerful. That's the goal with this window of opportunity.

> > high market penetration this will be the right micro for starting
> > Forth, unbeatable in price.
>
> That is an amazingly narrow basis for picking one processor family
> over another.  You picked one obscure, very specific configuration and
> compared the price point.  That tells you that for this one
> combination of features the ARM processor is not the best choice...
> so?  What about the other 2**256 other combinations of features?  If
> you leave out the Ethernet I don't think the PIC32 is the best choice
> anymore.  If you require more than 12 bits of resolution I believe
> that leaves the PIC32 out entirely!

You are totally right. But this was not the main reason for my quest
to put a Forth onto the PIC32.
My real reason was to bring the seed of Forth into this newbies
community.
Is there any other newbies community which is bigger or where it is
easier to step in?

> > > > In Australia I got a little box called "MAXIMITE" with this micro, and
> > > > it would be great to have Forth on it. Is anyone out there to help me
> > > > to get a MIPS Forth for the PIC32?
>
> > But reading the Gforth goals, I found these words: "It [Gforth] should
> > become standard, i.e. widely accepted and used. This goal is the most
> > difficult one." In my opinion the PIC32 is the window of opportunity
> > to reach this goal.
>
> I think you will find porting a Forth to the ARM processor will result
> in it be use much, much more!
>

Forth - in this case named Gforth - becoming a standard surely doesn't
mean it should be on M$ and the ARM only.

> > I installed Gforth on my laptop, it works really well.
> > And then the next surprise: the source code  includes a MIPS
> > assembler!!!
>
> > It would be great if somebody will take the job to make Gforth run on
> > the PIC32.
> > As I wrote, a window of opportunity. And most of the work seems to be
> > done.
>
> Please don't misunderstand me.  If someone wants to port gforth to the
> MIPS processor that is a great thing.  I am just pointing out that I
> think the reasons you state for picking the MIPS architecture are not
> completely accurate.  I personally would much prefer to see forths
> ported to the ARM architecture because of its pervasiveness.  I think
> in the grand scheme of things the differences between the processors
> themselves won't amount to a hill of beans.  But ARM MCUs are being
> sold by how many dozen vendors now?  The MIPS is still just what, one,
> two, maybe three?  If you want flexibility and selection in MCUs, the
> ARM is the way to go.

Sure you are right. But it would be a pity to miss this window of
opportunity.

> One other point.  Porting any PC oriented forth to an MCU is nothing
> like porting between operating systems.  What is needed I think is a
> Forth cross-compiler that targets a given MCU and I think that is a
> very different animal requiring a lot of work... if it wasn't I think
> we would already see at least as many cross-compilers as there are
> forths.  There is one for the ARM,http://pygmy.utoh.org/riscy/
>
> Rick

Thanks for this hint and all your very helpful arguments!

Dirk.

[toc] | [prev] | [next] | [standalone]

#1172

Bluebee <visualforth@rocketmail.com> writes:
> And to my surprise, there is no ARM microprocessor available
> at that low price with that high amount of memory (512k program, 128k
> RAM), UART/SPI/I2C, RTCC, 16 AD channels, 5 Timers/PWM, USB/CAN/
> ETHERNET, 8 DMA's, and only 64 pins. Microchip licensed the MIPS
> architecture, a really powerful machine, and because of Microchips
> high market penetration this will be the right micro for starting
> Forth, unbeatable in price.

Did you look at the TI MSP430 stuff (16 bit)?  They have some
ultra-cheap boards and should be good hosts for Forth.  They even have a
wristwatch with one of those processors, that would be a good Forth
target.  We had a thread about it a few months ago.  I actually got to
see one of the watches and put it on my wrist for a minute recently.
The thing weighs a ton compared to a normal watch, and the display is
kind of hard to read, but I guess it's not outlandish.

[toc] | [prev] | [next] | [standalone]

#1182

On 12 Apr., 03:17, Paul Rubin <no.em...@nospam.invalid> wrote:
> Bluebee <visualfo...@rocketmail.com> writes:
> > ... Microchip licensed the MIPS
> > architecture, a really powerful machine, and because of Microchips
> > high market penetration this will be the right micro for starting
> > Forth, unbeatable in price.
>
> Did you look at the TI MSP430 stuff (16 bit)?  They have some
> ultra-cheap boards and should be good hosts for Forth.  

I just finished a project using a MSP430F2012 because of low power
consumption and because of low price. When I was starting the project
the MSP430F2012 was the only MSP430 with DIL footprint. This project
is for a little battery powered price sensitive small market consumer
gadget - I have produced  more than one hundred now.

The MSP430F2012 micro has only 2k FLASH and 128 Byte RAM, so I had to
deal with C which I never liked to do, using copy/paste from samples
first to get started. To find out how "while" works, I had to download
Kernighan Ritchie. Using C gave me the feeling to be back in "stone
age" - that means back to thirty years ago where people wrote
microprocessor programs with trial and error instead of working on the
target itself using a monitor or Forth.

TI now has a new development system targeted at newbies, for only
$4.30 including all the development tools for programming in C or
assembler. There is a free Forth missing, but these new
microprocessors, called "value line" have only 8kB FLASH and 256 Byte
RAM, so Forth would occupy the whole program memory, I guess. Of
course, this would be a window of opportunity to launch a free Forth
(TI's MSP430 $4.30 development board is called "Launchpad"), but a
much harder task. It has to be a tethered Forth in this case.
Is there a free portable umbilical Forth available?

[toc] | [prev] | [next] | [standalone]

#1223

Bluebee <visualforth@rocketmail.com> writes:
>> Did you look at the TI MSP430 stuff (16 bit)?  They have some
>> ultra-cheap boards and should be good hosts for Forth.  
>...
> The MSP430F2012 micro has only 2k FLASH and 128 Byte RAM, so I had to
> deal with C which I never liked to do, using copy/paste from samples

There are a bunch of different models.  The Chrono wristwatch I
mentioned uses CC430F6137 which has 4k of ram and 32k of flash.  That
seems quite adequate for a forth system organized to put most of the
static words in flash.

[toc] | [prev] | [next] | [standalone]

#1225

On 16 Apr., 15:50, Paul Rubin <no.em...@nospam.invalid> wrote:
> Bluebee <visualfo...@rocketmail.com> writes:
> >> Did you look at the TI MSP430 stuff (16 bit)?  They have some
> >> ultra-cheap boards and should be good hosts for Forth.  
> >...
> > The MSP430F2012 micro has only 2k FLASH and 128 Byte RAM, so I had to
> > deal with C which I never liked to do, using copy/paste from samples
>
> There are a bunch of different models.  The Chrono wristwatch I
> mentioned uses CC430F6137 which has 4k of ram and 32k of flash.  That
> seems quite adequate for a forth system organized to put most of the
> static words in flash.

Sure. I know. I wrote that I have chosen "a MSP430F2012 because of low
power consumption and because of low price. When I was starting the
project the MSP430F2012 was the only MSP430 with DIL footprint. This
project is for a little battery powered price sensitive small market
consumer
gadget..."

The other MSP micros are only available in SMT. I am living in an
rural area in Southern Pennsylvania and here is no infrastructure
supporting SMT.

But there is hope: some days ago I got the brand new MSP430G2452IN20,
part of TI's Value Line, with 8kByte FLASH, 256 Byte RAM and DIL 20
pin footprint, I do not have to change my PCB layout to use it, and it
is even 44 Cents cheaper than the MSP430F2012.

The cheapest CC430F6137 is 9$ @100pcs, 6 times as much as the
MSP430G2452IN20. Its impossible to convince my customer to pay a lot
more for hardware because I like to use Forth for easier programming.

[toc] | [prev] | [next] | [standalone]

#1226

On Apr 16, 6:05 pm, Bluebee <visualfo...@rocketmail.com> wrote:
> The other MSP micros are only available in SMT. I am living in an
> rural area in Southern Pennsylvania and here is no infrastructure
> supporting SMT.

I'm not sure what that even means.  Are you telling me you can't find
a soldering iron in PA?  Many surface mount parts can be easily hand
soldered just as easily if not more so than DIPs.  I've even done it
and I have a hand tremor that can be so bad I can barely write
legibly!


> But there is hope: some days ago I got the brand new MSP430G2452IN20,
> part of TI's Value Line, with 8kByte FLASH, 256 Byte RAM and DIL 20
> pin footprint, I do not have to change my PCB layout to use it, and it
> is even 44 Cents cheaper than the MSP430F2012.
>
> The cheapest CC430F6137 is 9$ @100pcs, 6 times as much as the
> MSP430G2452IN20. Its impossible to convince my customer to pay a lot
> more for hardware because I like to use Forth for easier programming.

The other option is to use a part mounted on a PCB.  Someone in the
FPGA group mounted a Silicon Blue device (fine pitch QFN) on a small
PCB which uses half via holes as solder points along the edge of the
board so it can be mounted on other boards much more easily than the
QFN.  I'm not sure what he charges for it, but it's not much.

I guess I'm assuming you are not working with large volumes or you
would have figured out that it can be very inexpensive to have boards
built by pros.  I work with a company in Shippensburg called Niche.
That give me excellent support.  Are you anywhere near there?  I use
them and they are an hour drive from me.  Once I started using them I
even stopped buying my own parts.  They get much better pricing from
the vendors than I do and they really don't add much on top.  Turnkey
is a great way to go for any real quantity.

Rick

[toc] | [prev] | [next] | [standalone]

#1230

On 16 Apr., 18:34, rickman <gnu...@gmail.com> wrote:
> On Apr 16, 6:05 pm, Bluebee <visualfo...@rocketmail.com> wrote:
>
> > The other MSP micros are only available in SMT. I am living in an
> > rural area in Southern Pennsylvania and here is no infrastructure
> > supporting SMT.
>
> I guess I'm assuming you are not working with large volumes or you
> would have figured out that it can be very inexpensive to have boards
> built by pros.  I work with a company in Shippensburg called Niche.
> That give me excellent support.  Are you anywhere near there?  I use
> them and they are an hour drive from me.  Once I started using them I
> even stopped buying my own parts.  They get much better pricing from
> the vendors than I do and they really don't add much on top.  Turnkey
> is a great way to go for any real quantity.
>
> Rick

Rick,

Thanks a lot for your comment! I guess that's just what I needed.
Niche is only 2½ hours from here, that is really no distance for the
United States. Before I came over to the US after loosing my job at
Lucent Technologies (when they closed their manufacturing plants
worldwide), I worked self employed in Nuremberg, Germany, having all
infrastructure I needed within less than half an hours reach. Three
PCB manufacturers, and a little shop did machine soldering for me,
even SMT.

I restarted working self employed in PA, and I am glad that I got this
little gadget to develop and to produce. You are right, I never did
work with high volume. This gadget now is the highest volume and the
lowest priced one I ever had. Up to now I only produced 120 of it, all
hand soldered.

I got this project from somebody from China, who's company starts
manufacturing only if there are 10,000 pcs or more to produce, they
did look for somebody who could do small volume to satisfy the demand
of one of their customers in Texas. The end user price of the gadget
is 90$ - there are ten dollars for manufacturing it.

I looked at Niche Electronics website, and it looks like they have
automatic pick and place machines.
I am not sure if they would have been able to populate and solder 100
PCBs 2¼"x3" with 50 parts for 600$ including setup fee.

But anyway, thanks a lot for this information. This gives hope for
future projects. Of course I would prefer SMT, and up to now I didn't
know where to get a recommendation for a reliable and affordable
company doing the job. Now I have to find out to get this going,
because I already told my customer that I will change to SMT with
higher volume.

Thanks a lot!

Dirk.

[toc] | [prev] | [next] | [standalone]

#1231

On Apr 16, 7:37 pm, Bluebee <visualfo...@rocketmail.com> wrote:
> On 16 Apr., 18:34, rickman <gnu...@gmail.com> wrote:
>
>
>
> > On Apr 16, 6:05 pm, Bluebee <visualfo...@rocketmail.com> wrote:
>
> > > The other MSP micros are only available in SMT. I am living in an
> > > rural area in Southern Pennsylvania and here is no infrastructure
> > > supporting SMT.
>
> > I guess I'm assuming you are not working with large volumes or you
> > would have figured out that it can be very inexpensive to have boards
> > built by pros.  I work with a company in Shippensburg called Niche.
> > That give me excellent support.  Are you anywhere near there?  I use
> > them and they are an hour drive from me.  Once I started using them I
> > even stopped buying my own parts.  They get much better pricing from
> > the vendors than I do and they really don't add much on top.  Turnkey
> > is a great way to go for any real quantity.
>
> > Rick
>
> Rick,
>
> Thanks a lot for your comment! I guess that's just what I needed.
> Niche is only 2½ hours from here, that is really no distance for the
> United States. Before I came over to the US after loosing my job at
> Lucent Technologies (when they closed their manufacturing plants
> worldwide), I worked self employed in Nuremberg, Germany, having all
> infrastructure I needed within less than half an hours reach. Three
> PCB manufacturers, and a little shop did machine soldering for me,
> even SMT.
>
> I restarted working self employed in PA, and I am glad that I got this
> little gadget to develop and to produce. You are right, I never did
> work with high volume. This gadget now is the highest volume and the
> lowest priced one I ever had. Up to now I only produced 120 of it, all
> hand soldered.
>
> I got this project from somebody from China, who's company starts
> manufacturing only if there are 10,000 pcs or more to produce, they
> did look for somebody who could do small volume to satisfy the demand
> of one of their customers in Texas. The end user price of the gadget
> is 90$ - there are ten dollars for manufacturing it.
>
> I looked at Niche Electronics website, and it looks like they have
> automatic pick and place machines.
> I am not sure if they would have been able to populate and solder 100
> PCBs 2¼"x3" with 50 parts for 600$ including setup fee.
>
> But anyway, thanks a lot for this information. This gives hope for
> future projects. Of course I would prefer SMT, and up to now I didn't
> know where to get a recommendation for a reliable and affordable
> company doing the job. Now I have to find out to get this going,
> because I already told my customer that I will change to SMT with
> higher volume.
>
> Thanks a lot!
>
> Dirk.

No, $600 for 100 boards might be a bit hard to meet, but it depends on
the complexity.  I think they charged me $250 for setup, but that was
after the initial pick and place programming, I think.  I did a run of
4 prototypes before they assembled the other 100 boards on my last
revision and they added the $250.  This is a small board, 4.5" x 0.85"
with some 200 parts about.  A lot of the unit cost is time on the PnP
machine and will depend on the number of parts.  Of course, DIP is not
SMT friendly.  I'm not sure they even have a solder wave machine to do
DIP type stuff.  But give them a call, they can be very
accommodating.  The big problem for you is that they will all need to
be built at one time.  It sounds like you need them a few at a time.

Where exactly are you?  There may be others closer.  A vendor wants me
to work with Win-somebody just north of Pittsburgh.  That's five hours
from me, but if the vendor gets me the work, I'll use his assembly
house to return the favor.

Rick

[toc] | [prev] | [next] | [standalone]

#1233

On 16 Apr., 19:55, rickman <gnu...@gmail.com> wrote:
> On Apr 16, 7:37 pm, Bluebee <visualfo...@rocketmail.com> wrote:
>
> No, $600 for 100 boards might be a bit hard to meet, but it depends on
> the complexity.  I think they charged me $250 for setup, but that was
> after the initial pick and place programming, I think.  

$250 is reasonable. The same setup fee I pay for PCBs, that's why I
always order 100 PCBs minimum. Seems to be you had the same quantity!
But I always get only orders of thirty gadgets each, that may make it
impossible to use SMT.

> ... Of course, DIP is not
> SMT friendly.  

Yes, not all parts are available with SMT. I need three special audio
jacks to fit into the little box, and two Keystone contacts for the
battery. Everything else may be SMDs. I guess the special through hole
parts will make it expensive then.

> ... But give them a call, they can be very
> accommodating. The big problem for you is that they will all need to
> be built at one time.  It sounds like you need them a few at a time.

I guess I have to wait for a fitting project!

> Where exactly are you?  There may be others closer.  A vendor wants me
> to work with Win-somebody just north of Pittsburgh.  That's five hours
> from me, but if the vendor gets me the work, I'll use his assembly
> house to return the favor.
>
I am living in Somerset county, and Pittsburgh is 2½ hours away from
here, too.

Nice to chat with somebody who is really working on electronics, too!
I appreciate your help very much!

Dirk.

[toc] | [prev] | [next] | [standalone]

#1271

On Apr 16, 8:41 pm, Bluebee <visualfo...@rocketmail.com> wrote:
> On 16 Apr., 19:55, rickman <gnu...@gmail.com> wrote:
>
> > On Apr 16, 7:37 pm, Bluebee <visualfo...@rocketmail.com> wrote:
>
> > No, $600 for 100 boards might be a bit hard to meet, but it depends on
> > the complexity.  I think they charged me $250 for setup, but that was
> > after the initial pick and place programming, I think.  
>
> $250 is reasonable. The same setup fee I pay for PCBs, that's why I
> always order 100 PCBs minimum. Seems to be you had the same quantity!
> But I always get only orders of thirty gadgets each, that may make it
> impossible to use SMT.

No, not impossible, but it will run the cost up some.  Or you can
build 100 if you know the orders will come.


> > ... Of course, DIP is not
> > SMT friendly.  
>
> Yes, not all parts are available with SMT. I need three special audio
> jacks to fit into the little box, and two Keystone contacts for the
> battery. Everything else may be SMDs. I guess the special through hole
> parts will make it expensive then.

No, a few through hole parts don't run the cost up much.  Even though
my board is all SMT, they hand place a couple of connectors.  Not sure
why, but I was stuck with the positioning since it has to be
compatible with an existing mother board it mounts to so the connector
body over hangs the board a touch.


> > ... But give them a call, they can be very
> > accommodating. The big problem for you is that they will all need to
> > be built at one time.  It sounds like you need them a few at a time.
>
> I guess I have to wait for a fitting project!
>
> > Where exactly are you?  There may be others closer.  A vendor wants me
> > to work with Win-somebody just north of Pittsburgh.  That's five hours
> > from me, but if the vendor gets me the work, I'll use his assembly
> > house to return the favor.
>
> I am living in Somerset county, and Pittsburgh is 2½ hours away from
> here, too.
>
> Nice to chat with somebody who is really working on electronics, too!
> I appreciate your help very much!

It looks like you are about 90 miles to Shippensburg, not counting all
the curves in the roads.  There is also an outfit in Hagerstown,
Tristate I believe which is around the same distance.  I haven't
worked with Tristate.  The first quote I got from them was a bit high
and I ended up with Niche.  I like them so much I don't even compete
it any more.

I seem to recall an outfit in Altoona, Kuchera, IIRC.  That is even
closer to you.  You are also not far from Cumberland which may have an
assembly house.

BTW, I found Sommerset on the map, how can you be 2 1/2 hours from
Pittsburgh if it is only 60 miles?  Are the roads that tricky?

Reminds me of the line from "O' Brother", "A geographical oddity, 2
weeks from everywhere".

Rick

[toc] | [prev] | [next] | [standalone]

Page 1 of 3  [1] 2 3  Next page →

Back to top | Article view | comp.lang.forth

csiph-web