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

Groups > comp.lang.forth > #26055 > unrolled thread

Id love to surf web with 4megs ram forth pc using forth and forth chips

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

Contents

  Id love to surf web with 4megs ram forth pc using forth and forth chips the_gavino_himself <visphatesjava@gmail.com> - 2013-09-27 14:52 -0700
    Re: Id love to surf web with 4megs ram forth pc using forth and forth chips Elizabeth D Rather <erather@forth.com> - 2013-09-27 12:35 -1000
      Re: Id love to surf web with 4megs ram forth pc using forth and forth chips Bernd Paysan <bernd.paysan@gmx.de> - 2013-10-02 00:01 +0200
        Re: Id love to surf web with 4megs ram forth pc using forth and forth chips Mark Wills <markrobertwills@yahoo.co.uk> - 2013-10-02 00:35 -0700
          Re: Id love to surf web with 4megs ram forth pc using forth and forth chips rickman <gnuarm@gmail.com> - 2013-10-02 04:00 -0400
        Re: Id love to surf web with 4megs ram forth pc using forth and forth chips Paul Rubin <no.email@nospam.invalid> - 2013-10-02 01:34 -0700
          Re: Id love to surf web with 4megs ram forth pc using forth and forth chips Mark Wills <markrobertwills@yahoo.co.uk> - 2013-10-02 07:30 -0700
        Re: Id love to surf web with 4megs ram forth pc using forth and forth chips albert@spenarnc.xs4all.nl (Albert van der Horst) - 2013-10-02 10:26 +0000
          Re: Id love to surf web with 4megs ram forth pc using forth and forth chips Bill Richards <billrich@gmx.com> - 2013-10-02 14:41 +0000
            Re: Id love to surf web with 4megs ram forth pc using forth and forth chips Mark Wills <markrobertwills@yahoo.co.uk> - 2013-10-02 08:25 -0700
              Re: Id love to surf web with 4megs ram forth pc using forth and forth chips rickman <gnuarm@gmail.com> - 2013-10-02 23:13 -0400
          Re: Id love to surf web with 4megs ram forth pc using forth and forth chips Matthias Koch <matthias.koch@hot.uni-hannover.de> - 2013-10-25 12:46 +0200
        Re: Id love to surf web with 4megs ram forth pc using forth and forth chips the_gavino_himself <visphatesjava@gmail.com> - 2013-11-06 17:00 -0800
    Re: Id love to surf web with 4megs ram forth pc using forth and forth chips Brad Eckert <hwfwguy@gmail.com> - 2013-10-04 08:09 -0700
      Re: Id love to surf web with 4megs ram forth pc using forth and forth chips "Ed" <invalid@invalid.com> - 2013-10-05 11:45 +1000
      Re: Id love to surf web with 4megs ram forth pc using forth and forth chips gavino_himself <visploveslisp@gmail.com> - 2013-11-17 13:55 -0800
    Re: Id love to surf web with 4megs ram forth pc using forth and forth chips "Clyde W. Phillips Jr." <cwpjr02@gmail.com> - 2013-10-15 16:45 -0700
      Re: Id love to surf web with 4megs ram forth pc using forth and forth chips Spam@ControlQ.com - 2013-10-16 11:20 -0400
        Re: Id love to surf web with 4megs ram forth pc using forth and forth chips albert@spenarnc.xs4all.nl (Albert van der Horst) - 2013-10-17 12:03 +0000
          Re: Id love to surf web with 4megs ram forth pc using forth and forth chips Spam@ControlQ.com - 2013-10-17 11:40 -0400
            Re: Id love to surf web with 4megs ram forth pc using forth and forth chips albert@spenarnc.xs4all.nl (Albert van der Horst) - 2013-10-17 19:58 +0000
              Re: Id love to surf web with 4megs ram forth pc using forth and forth chips Spam@ControlQ.com - 2013-10-17 17:22 -0400
                Re: Id love to surf web with 4megs ram forth pc using forth and forth chips "Elizabeth D. Rather" <erather@forth.com> - 2013-10-17 11:45 -1000
                  Re: Id love to surf web with 4megs ram forth pc using forth and forth chips Spam@ControlQ.com - 2013-10-18 13:33 -0400
                    Re: Id love to surf web with 4megs ram forth pc using forth and forth chips gavino_himself <visploveslisp@gmail.com> - 2013-11-17 14:15 -0800
              Re: Id love to surf web with 4megs ram forth pc using forth and forth chips "Rod Pemberton" <dont_use_email@nohavenotit.com> - 2013-10-18 04:02 -0400
                Re: Id love to surf web with 4megs ram forth pc using forth and forth chips Stefan Mauerhofer <smauerhofer@androsoft.ch> - 2013-10-18 07:01 -0700
                  Re: Id love to surf web with 4megs ram forth pc using forth and forth chips "Elizabeth D. Rather" <erather@forth.com> - 2013-10-18 08:53 -1000
                    Re: Id love to surf web with 4megs ram forth pc using forth and forth chips c-gschuette@neogov.net - 2013-11-07 09:24 -0800
                      Re: Id love to surf web with 4megs ram forth pc using forth and forth chips "Elizabeth D. Rather" <erather@forth.com> - 2013-11-07 07:38 -1000
                      Re: Id love to surf web with 4megs ram forth pc using forth and forth chips "Rod Pemberton" <dont_use_email@xnohavenotit.cnm> - 2013-11-07 20:07 -0500
                        Re: Id love to surf web with 4megs ram forth pc using forth and forth chips the_gavino_himself <visphatesjava@gmail.com> - 2013-11-16 10:00 -0800
                          Re: Id love to surf web with 4megs ram forth pc using forth and forth chips "Rod Pemberton" <dont_use_email@xnohavenotit.cnm> - 2013-11-16 17:22 -0500
                            Re: Id love to surf web with 4megs ram forth pc using forth and forth chips "Elizabeth D. Rather" <erather@forth.com> - 2013-11-16 13:53 -1000
                              Re: Id love to surf web with 4megs ram forth pc using forth and forth chips Andrew Haley <andrew29@littlepinkcloud.invalid> - 2013-11-17 04:17 -0600
                                Re: Id love to surf web with 4megs ram forth pc using forth and forth chips Bernd Paysan <bernd.paysan@gmx.de> - 2013-11-17 15:21 +0100
                                Re: Id love to surf web with 4megs ram forth pc using forth and forth chips "Elizabeth D. Rather" <erather@forth.com> - 2013-11-17 08:55 -1000
                                  Re: Id love to surf web with 4megs ram forth pc using forth and forth chips Andrew Haley <andrew29@littlepinkcloud.invalid> - 2013-11-17 16:13 -0600
                                    Re: Id love to surf web with 4megs ram forth pc using forth and forth chips "Elizabeth D. Rather" <erather@forth.com> - 2013-11-17 12:37 -1000
                  Re: Id love to surf web with 4megs ram forth pc using forth and forth chips "Rod Pemberton" <dont_use_email@nohavenotit.com> - 2013-10-18 21:19 -0400
                  Re: Id love to surf web with 4megs ram forth pc using forth and forth chips gavino_himself <visploveslisp@gmail.com> - 2013-11-17 14:28 -0800
                  Re: Id love to surf web with 4megs ram forth pc using forth and forth chips gavino_himself <visploveslisp@gmail.com> - 2013-11-17 14:34 -0800
          Re: Id love to surf web with 4megs ram forth pc using forth and forth chips "Elizabeth D. Rather" <erather@forth.com> - 2013-10-17 11:32 -1000
          Re: Id love to surf web with 4megs ram forth pc using forth and forth chips "Rod Pemberton" <dont_use_email@nohavenotit.com> - 2013-10-18 04:02 -0400
        Re: Id love to surf web with 4megs ram forth pc using forth and forth chips gavino_himself <visploveslisp@gmail.com> - 2013-11-17 14:11 -0800

Page 2 of 3 — ← Prev page 1 [2] 3  Next page →

#26617

In article <alpine.BSF.2.00.1310171128430.51137@yoko.controlq.com>,
 <Spam@ControlQ.com> wrote:
>
>On Thu, 17 Oct 2013, Albert van der Horst wrote:
>
>>> I disagree with the big $$$ assertion.  Linux started as a student
>>> project, and has evolved to where it is today.  A generic Forth OS with a
>>> degree of portability can be had quite reasonably, even as an open source
>>> hobby-type project.
>>
>> This misrepresents the situation. Torvalds jumped on the GNU-train that
>> was running 200 km/h and had consumed hundreds of millions of dollars in
>> public funding (well spent, IMO).
>
>True, a readily available C compiler was a pre-requisite for Linux, but a
>Forth based X-assembler for a given target fulfills that need.  In
>addition, knowledge of the boot code/initialization for the target.  That
>does not constitute a huge portion of the OS, but would cover most of the
>machine dependencies.

Torvalds need not worry about You forget bash lex yacc awk cp ls passwd man info zip mount nroff mail ping
at cron

Must I go on? That is all part of GNU. A Forth based X-assembler gives you
nothing of those.

>
>>
>>>
>>> The one thing that would simplify the portability is to use a device tree
>>> to describe the target, in a manner similar to the Flattened Device Tree
>>> model used by the PPC folks.  This will describe memory, registers, GPIO,
>>> i2C buses and various components of the board, as well as the SoC itself
>>> to produce a "bus model".  Drivers can then be developed generically.
>>
>> We don't have to start from nothing. There is open boot, and "all" we have
>> to do is to port that to Pentium machines.
>
>Assuming this belongs on a desktop.  Arm, PPC and MIPS would be better
>first targets ... and again, I would stress the FDT approach to describe
>the platform/architecture to the OS -- the modern PC's firmware with bus
>instantiation is complex, but is not found on embedded targets.
>
>>>
>>> Basically, one would implement basic boot code in assembler, plus the
>>> primitives (Ting has suggested somewhere between 20 and 30 or so), and the
>>> rest would be written in Forth.
>>
>> The rest... Have you looked at an obscure language like Python recently?
>> No OS written in Python. Still getting an overview of what
>> libraries are available is mind-boggling.
>
>Python is a general purpose high level scripting language with a degree of
>object orientation, and decent performance.  It naturally attracts coders
>where Forth does its best to drive them away.  I'm not sure what your
>point is here, other than Python is a wildly successful scripting language
>whereas Forth is relegated to fringe status.

My point is the amount of libraries needed to start an OS is mind-boggling.
It takes a Gavino to casually propose things like this.

Groetjes Albert
-- 
Albert van der Horst, UTRECHT,THE NETHERLANDS
Economic growth -- being exponential -- ultimately falters.
albert@spe&ar&c.xs4all.nl &=n http://home.hccnet.nl/a.w.m.van.der.horst

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

#26619

On Thu, 17 Oct 2013, Albert van der Horst wrote:

> Date: 17 Oct 2013 19:58:04 GMT
> From: Albert van der Horst <albert@spenarnc.xs4all.nl>
> Newsgroups: comp.lang.forth
> Subject: Re: Id love to surf web with 4megs ram forth pc using forth and forth
>      chips
> 
> In article <alpine.BSF.2.00.1310171128430.51137@yoko.controlq.com>,
> <Spam@ControlQ.com> wrote:
>>
>> On Thu, 17 Oct 2013, Albert van der Horst wrote:
>>
>>>> I disagree with the big $$$ assertion.  Linux started as a student
>>>> project, and has evolved to where it is today.  A generic Forth OS with a
>>>> degree of portability can be had quite reasonably, even as an open source
>>>> hobby-type project.
>>>
>>> This misrepresents the situation. Torvalds jumped on the GNU-train that
>>> was running 200 km/h and had consumed hundreds of millions of dollars in
>>> public funding (well spent, IMO).
>>
>> True, a readily available C compiler was a pre-requisite for Linux, but a
>> Forth based X-assembler for a given target fulfills that need.  In
>> addition, knowledge of the boot code/initialization for the target.  That
>> does not constitute a huge portion of the OS, but would cover most of the
>> machine dependencies.
>
> Torvalds need not worry about You forget bash lex yacc awk cp ls passwd man info zip mount nroff mail ping
> at cron
>
> Must I go on? That is all part of GNU. A Forth based X-assembler gives you
> nothing of those.

Hosting your X-dev environment on Unix/Linux gives you all of that.  Are 
you seriously proposing to re-implement all of [U|Li]nux in Forth?

Trying to re-implement Linux is pointless (we already have it). 
Implementing a small but useful RTOS/Boot environment is potentially 
useful -- but shooting for full posix compliance is a fool's errand 
(thus explaining Gavino's interest in doing so).

>> Python is a general purpose high level scripting language with a degree of
>> object orientation, and decent performance.  It naturally attracts coders
>> where Forth does its best to drive them away.  I'm not sure what your
>> point is here, other than Python is a wildly successful scripting language
>> whereas Forth is relegated to fringe status.
>
> My point is the amount of libraries needed to start an OS is mind-boggling.
> It takes a Gavino to casually propose things like this.
>
> Groetjes Albert
>

Your use of the term "library" reveals a bias, methinks, and possibly 
some "in the box" thinking.  This surprises me given your recent foray 
into re-doing JonesForth ... for fun, I presume.

Complexity depends upon the OS.  API's for scheduling, interrupt handling 
and simple process creation/suspension/preemption/deletion don't take 
much.  The bigger chore is device drivers for I2C, flash, serial i/o 
(usb/uart) etc., and that is significantly helped by the use of device 
trees to describe the resources.

It easy to say "no" to a non-trivial project, and much harder to find the 
rationale to proceed.  Personally, I look forward to the day when GavinOS 
(TM) is available, and running FireFox in emulation on multiple platforms. 8-)

Cheers,
Rob.

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

#26621

On 10/17/13 11:22 AM, Spam@ControlQ.com wrote:
> On Thu, 17 Oct 2013, Albert van der Horst wrote:
>
...
>>> Forth based X-assembler for a given target fulfills that need.  In
>>> addition, knowledge of the boot code/initialization for the target.
>>> That
>>> does not constitute a huge portion of the OS, but would cover most of
>>> the
>>> machine dependencies.
>>
>> Torvalds need not worry about You forget bash lex yacc awk cp ls
>> passwd man info zip mount nroff mail ping
>> at cron
>>
>> Must I go on? That is all part of GNU. A Forth based X-assembler gives
>> you
>> nothing of those.
>
> Hosting your X-dev environment on Unix/Linux gives you all of that.  Are
> you seriously proposing to re-implement all of [U|Li]nux in Forth?
>
> Trying to re-implement Linux is pointless (we already have it).
> Implementing a small but useful RTOS/Boot environment is potentially
> useful -- but shooting for full posix compliance is a fool's errand
> (thus explaining Gavino's interest in doing so).

Exactly. The first thing you have to do is define exactly what this 
project is, what market it's intended to serve, and what capabilities it 
will need to serve that market. This is far from clear right now!

Cheers,
Elizabeth

-- 
==================================================
Elizabeth D. Rather   (US & Canada)   800-55-FORTH
FORTH Inc.                         +1 310.999.6784
5959 West Century Blvd. Suite 700
Los Angeles, CA 90045
http://www.forth.com

"Forth-based products and Services for real-time
applications since 1973."
==================================================

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

#26634

On Thu, 17 Oct 2013, Elizabeth D. Rather wrote:

>> Hosting your X-dev environment on Unix/Linux gives you all of that.  Are
>> you seriously proposing to re-implement all of [U|Li]nux in Forth?
>> 
>> Trying to re-implement Linux is pointless (we already have it).
>> Implementing a small but useful RTOS/Boot environment is potentially
>> useful -- but shooting for full posix compliance is a fool's errand
>> (thus explaining Gavino's interest in doing so).
>
> Exactly. The first thing you have to do is define exactly what this project 
> is, what market it's intended to serve, and what capabilities it will need to 
> serve that market. This is far from clear right now!
>
> Cheers,
> Elizabeth

Cooler heads prevail!

I would point at the remarkable FreeRTOS project for scoping ideas. 
Better yet, a full featured forth which runs on FreeRTOS might just be the 
ticket ... This alludes to Albert's suggestion that one might gain an 
advantage by standing on the shoulders of giants, though his point 
referred to the Gnu tool chain and libraries.  It also focuses upon 
Forth's strengths in the embedded domain.

One might hope also for this project to be actually useful, that a 
permissive license be used, and NOT GPL, but who knows what Gavino might 
do in order to run FireCat 8-)??

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

#26871

at the end you simply want a web broswer

converely a fast webserver

then you have the client server model that makes the net

if forth can replace bsd ro linux and do it simpler and well then I woudl lvoe to see it and would even try and help

once the platform is in place for network and web browser let the community come up with other stuf

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

#26629

On Thu, 17 Oct 2013 15:58:04 -0400, Albert van der Horst  
<albert@spenarnc.xs4all.nl> wrote:

> My point is the amount of libraries needed to start an OS is  
> mind-boggling.

You just need a full featured C compiler to code an OS in C.
That's it.  30% or so of the C language is independent of
the host and another 40% is custom to each host, but works
the same across different platforms.  30% is not portable.
Of course, a full featured C compiler provides inline assembly
which is needed on the x86 platform to provide special
x86 processor instructions.  However, no C libraries
are needed to bootstrap an OS in C.  A C library isn't that
dependent on an OS either.  To build a C library, you only
need about 18 to 20 functions provided by the OS, plus a
large amount of C code to implement the library.  Of course,
an efficient C library can use over 200 functions provided
by the OS.

I'm not sure what the exact equivalent would be for Forth,
but most likely it's just the base Forth language plus some
assembly, perhaps meta-compilation (cross-compiling...) too.

If you're interested in bootstrapping an OS, come on over
to news://alt.os.development or http://wiki.osdev.org.

But, be forewarned that most guys either code an OS in
x86 assembly or C.  Very few use C++.  Very few develop
for non-x86.  I've never heard of Forth being used, probably
because Forth either is an OS already or isn't, depending
on the implementation.


Rod Pemberton
-- 
84Nk5y once again, proving artists, citizens, authorities from around
the world CAN'T recognize ART when they SEE it.  Rats, chimps, gays,
un-angelic angels, shell-shocked soldiers, distopian childhoods ...
Whats next?  We have Youtube where it's REAL, in COLOR, without need
of monochrome stenciled spray paint ... BANKSY, the world is unchanged.

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

#26633

> But, be forewarned that most guys either code an OS in
> 
> x86 assembly or C.  Very few use C++.  Very few develop
> 
> for non-x86.  I've never heard of Forth being used, probably
> 
> because Forth either is an OS already or isn't, depending
> 
> on the implementation.

There is no C required to run a machine. I've implemented a 64 bit (native Forth) operating system for the PC using assembler, Forth-like macros and Forth only. 

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

#26636

On 10/18/13 4:01 AM, Stefan Mauerhofer wrote:
>
>> But, be forewarned that most guys either code an OS in
>>
>> x86 assembly or C.  Very few use C++.  Very few develop
>>
>> for non-x86.  I've never heard of Forth being used, probably
>>
>> because Forth either is an OS already or isn't, depending
>>
>> on the implementation.
>
> There is no C required to run a machine. I've implemented a 64 bit (native Forth) operating system for the PC using assembler, Forth-like macros and Forth only.
>

Greg Bailey has supported such a system for many years, as a file & 
email server for companies with >1,000 users.

Greg's native Forth OSs are specifically tailored to their applications, 
however, and I suspect yours may be as well. Opening up such a system as 
a general purpose OS capable of running random user programs written in 
whatever language with a full web surfer would still be a non-trivial 
project that needs to be carefully designed.

Cheers,
Elizabeth

-- 
==================================================
Elizabeth D. Rather   (US & Canada)   800-55-FORTH
FORTH Inc.                         +1 310.999.6784
5959 West Century Blvd. Suite 700
Los Angeles, CA 90045
http://www.forth.com

"Forth-based products and Services for real-time
applications since 1973."
==================================================

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

#26808

well al u need is working web browser and some kinda way to save files....with tcp/ip and drivers for usb keyboard n mouse

try for windows 3.1 and netscape

in 1% code fo course

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

#26809

On 11/7/13 7:24 AM, c-gschuette@neogov.net wrote:
> well al u need is working web browser and some kinda way to save files....with tcp/ip and drivers for usb keyboard n mouse
>
> try for windows 3.1 and netscape
>
> in 1% code fo course
>

And what would be the market for this 30-yr-old technology?

Cheers,
Elizabeth

-- 
==================================================
Elizabeth D. Rather   (US & Canada)   800-55-FORTH
FORTH Inc.                         +1 310.999.6784
5959 West Century Blvd. Suite 700
Los Angeles, CA 90045
http://www.forth.com

"Forth-based products and Services for real-time
applications since 1973."
==================================================

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

#26811

On Thu, 07 Nov 2013 12:24:25 -0500, <c-gschuette@neogov.net> wrote:

> well al u need is working web browser and some kinda way to save  
> files....with tcp/ip and drivers for usb keyboard n mouse
>
> try for windows 3.1 and netscape
>
> in 1% code fo course

Well Gavino (a.k.a. Gavin Schuette),

That wouldn't be a "Forth PC using Forth and Forth chips"
would it?

I'm amazed you just recommend using a non-Forth solution.
You've rejected that idea from others repeatedly, including
me, over the course of a number of years.  Are you just
being the internet "pest" that numerous other Usenet groups,
IRC channels, online forums, claim you are?  Is that even
your real name?  Did you hack that IP?


Rod Pemberton

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

#26856

What idea was rejected?

To surf the net without windows mac or linux/bsd would be fun esp on 1% the code or less!!!

and especily on forth cpu!

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

#26857

On Sat, 16 Nov 2013 13:00:41 -0500, the_gavino_himself  
<visphatesjava@gmail.com> wrote:

> What idea was rejected?
>

You seemed to reject the idea of a "Forth PC using Forth
and Forth chips" by suggesting one should use "windows 3.1
and netscape" instead.  Why would you suggest that if you
hadn't rejected a Forth PC?

> To surf the net without windows mac or linux/bsd would
> be fun esp on 1% the code or less!!!
>

I equate the 1% of the code or less comment to mean slow,
since minimal code like that is only possible with threaded
code, which means an interpreted Forth.  I honestly don't
see how this is any different from using a Commodore 64 or
Timex Sinclair 1000 to browse the web.

If you're using a compiled Forth, the compiled code will
be about the same size as compiled C, likely slightly larger
since not as much work has been done to optimize Forth as
compare to C.  The operations must be converted to the host
processor's assembly.  If the code is well written and highly
optimized, the assembly code is going to result in near
identical code for both, and therefore be the same size
for both.  Why?  You can't achieve a greater efficiency
than the most optimal solution for that processor.  If
both C and Forth can generate the same code, why bother
with Forth?  Ditto Fortran, Pascal, Perl, Lisp, BASIC...

> and especily on forth cpu!

First, please learn the difference between a CPU and
microprocessor.  You mean microprocessor, sometimes
abbreviated with the lowercase greek letter mu and 'p'.
Of course, most keyboards don't have mu, but have 'u',
so sometimes posted as 'up', e.g., "forth up".

Second, that can improve speed, but can it improve speed
enough?  Intel, AMD, ARM, etc are *not* producing Forth
microprocessors.  So, at best, you have much, much older
microprocessor technology being used, unless Chuck Moore
is involved.  In which case, the processor is comparable
to other modern non-Forth designs, but usually no better.
If C.M. is not involved, the performance is much worse,
comparatively.

Given all of that above, what is the point?  To me,
it seems you don't really think much about the claims
you make.  You just seem to say: "It's cool!"  So, it's
cool, but it's also irrelevant and useless.  So, why
bother?  If it's just a novelty, who cares?  If you do,
make and sell some "do nothing" machines.


Rod Pemberton

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

#26858

On 11/16/13 12:22 PM, Rod Pemberton wrote:
> On Sat, 16 Nov 2013 13:00:41 -0500, the_gavino_himself
> <visphatesjava@gmail.com> wrote:
>
...
>> To surf the net without windows mac or linux/bsd would
>> be fun esp on 1% the code or less!!!
>>
>
> I equate the 1% of the code or less comment to mean slow,
> since minimal code like that is only possible with threaded
> code, which means an interpreted Forth.  I honestly don't
> see how this is any different from using a Commodore 64 or
> Timex Sinclair 1000 to browse the web.

Actually, optimized, compiled Forth turns out to be smaller than 
threaded code.

> If you're using a compiled Forth, the compiled code will
> be about the same size as compiled C, likely slightly larger
> since not as much work has been done to optimize Forth as
> compare to C.  The operations must be converted to the host
> processor's assembly.  If the code is well written and highly
> optimized, the assembly code is going to result in near
> identical code for both, and therefore be the same size
> for both.  Why?  You can't achieve a greater efficiency
> than the most optimal solution for that processor.  If
> both C and Forth can generate the same code, why bother
> with Forth?  Ditto Fortran, Pascal, Perl, Lisp, BASIC...

If it's well-factored, it will be smaller, often much smaller, with 
about the same performance. The tradeoffs are different because 
subroutine calls are considerably more expensive (in both size and 
microseconds) in C and other languages because of the cost of managing 
the calling sequence. Therefore, well-factored Forth will have a lot 
more calls and less repeated code.

Cheers,
Elizabeth

-- 
==================================================
Elizabeth D. Rather   (US & Canada)   800-55-FORTH
FORTH Inc.                         +1 310.999.6784
5959 West Century Blvd. Suite 700
Los Angeles, CA 90045
http://www.forth.com

"Forth-based products and Services for real-time
applications since 1973."
==================================================

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

#26864

Elizabeth D. Rather <erather@forth.com> wrote:
> 
> Actually, optimized, compiled Forth turns out to be smaller than 
> threaded code.

Hmm.  Optimized compiled Forth can be smaller than, say, 32-bit
threaded code, but it's not smaller than all forms of threaded code,
some of which can be extremely dense.

> On 11/16/13 12:22 PM, Rod Pemberton wrote:
>> If you're using a compiled Forth, the compiled code will be about
>> the same size as compiled C, likely slightly larger since not as
>> much work has been done to optimize Forth as compare to C.  The
>> operations must be converted to the host processor's assembly.  If
>> the code is well written and highly optimized, the assembly code is
>> going to result in near identical code for both,

That's true if you solve problems in the same way in both languages,
but IME you probably don't: Forth and C encourage different problem-
solving styles, in particular the way that programs are factored, so
the solutions aren't going to be the same.  It's hard to define this
in a precise way, but IMO there is such a thing as Forthly design.

>> and therefore be the same size for both.  Why?  You can't achieve a
>> greater efficiency than the most optimal solution for that
>> processor.  If both C and Forth can generate the same code, why
>> bother with Forth?

But if both Language X and Forth can generate the same code, why
bother with X?

Andrew.

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

#26865

Andrew Haley wrote:
> That's true if you solve problems in the same way in both languages,
> but IME you probably don't: Forth and C encourage different problem-
> solving styles, in particular the way that programs are factored, so
> the solutions aren't going to be the same.  It's hard to define this
> in a precise way, but IMO there is such a thing as Forthly design.

One key aspect of why Forth programs are often much smaller than programs in 
other languages trying to solve a similar problem is the "redefine the 
problem" approach. The problem in many IT projects is not very well-defined, 
and therefore, people tend to build in a lot of what Brooks calls 
"accidental complexity" into their products. Different languages stipulate 
different ways to deal with complexity.

Java programmers tend to write adapter classes to proxy factories, or any 
other combination of "design pattern" they have learned. Abstractions are 
seen as a value of their own, so they are build in the midth of free air.  
The programs are somewhat maintainable, but large and slow.  The real work 
is often delegated to a C++ native function.  Each layer of abstraction adds 
bugs to the system.

Fortran programmers (and you can write Fortran programs in any language) 
tend to write long linear stuff that does one thing after the other, 
everything broken down to the primitives of the language.  The programs are 
unmaintainable, large, but usually quite fast.  The real work is delegated 
to a library written 40 years ago.  Each change adds bugs to the system, 
which means the whole program is very fragile.

One should mention that every programmer spends more time debugging his 
stuff than writing.  Ease of debugging and trying different approaches is a 
strength rarely found.  Forth's style is to do only what's needed, and to do 
experiments - both with the code and with the spec.

-- 
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://bernd-paysan.de/

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

#26866

On 11/17/13 12:17 AM, Andrew Haley wrote:
> Elizabeth D. Rather <erather@forth.com> wrote:
>>
>> Actually, optimized, compiled Forth turns out to be smaller than
>> threaded code.
>
> Hmm.  Optimized compiled Forth can be smaller than, say, 32-bit
> threaded code, but it's not smaller than all forms of threaded code,
> some of which can be extremely dense.
>

When we first developed our optimizing compilers for SwiftX we were 
shocked to find that even on platforms such as the 8051 things turned 
out to be not much bigger and even sometimes smaller. One reason is that 
some things just go away: a multi-cell NEXT replaced by a simple RET, 
for example; no code field; things like that. Obviously some 
microcontrollers fare better than others.

Cheers,
Elizabeth

-- 
==================================================
Elizabeth D. Rather   (US & Canada)   800-55-FORTH
FORTH Inc.                         +1 310.999.6784
5959 West Century Blvd. Suite 700
Los Angeles, CA 90045
http://www.forth.com

"Forth-based products and Services for real-time
applications since 1973."
==================================================

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

#26870

Elizabeth D. Rather <erather@forth.com> wrote:
> On 11/17/13 12:17 AM, Andrew Haley wrote:
>> Elizabeth D. Rather <erather@forth.com> wrote:
>>>
>>> Actually, optimized, compiled Forth turns out to be smaller than
>>> threaded code.
>>
>> Hmm.  Optimized compiled Forth can be smaller than, say, 32-bit
>> threaded code, but it's not smaller than all forms of threaded code,
>> some of which can be extremely dense.
> 
> When we first developed our optimizing compilers for SwiftX we were 
> shocked to find that even on platforms such as the 8051 things turned 
> out to be not much bigger and even sometimes smaller. One reason is that 
> some things just go away: a multi-cell NEXT replaced by a simple RET, 
> for example; no code field; things like that. Obviously some 
> microcontrollers fare better than others.

I accept that, but you can do much tighter than 16-bit threaded code
on an 8051: for some space-constrained applications 8-bit
token-threading is a good choice because it doesn't much slow things
down (over,say, DTC) and saves a good deal of space.  I have
implemented this, and I think it's very hard to beat for compactness.

Andrew.

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

#26875

On 11/17/13 12:13 PM, Andrew Haley wrote:
> Elizabeth D. Rather <erather@forth.com> wrote:
>> On 11/17/13 12:17 AM, Andrew Haley wrote:
>>> Elizabeth D. Rather <erather@forth.com> wrote:
>>>>
>>>> Actually, optimized, compiled Forth turns out to be smaller than
>>>> threaded code.
>>>
>>> Hmm.  Optimized compiled Forth can be smaller than, say, 32-bit
>>> threaded code, but it's not smaller than all forms of threaded code,
>>> some of which can be extremely dense.
>>
>> When we first developed our optimizing compilers for SwiftX we were
>> shocked to find that even on platforms such as the 8051 things turned
>> out to be not much bigger and even sometimes smaller. One reason is that
>> some things just go away: a multi-cell NEXT replaced by a simple RET,
>> for example; no code field; things like that. Obviously some
>> microcontrollers fare better than others.
>
> I accept that, but you can do much tighter than 16-bit threaded code
> on an 8051: for some space-constrained applications 8-bit
> token-threading is a good choice because it doesn't much slow things
> down (over,say, DTC) and saves a good deal of space.  I have
> implemented this, and I think it's very hard to beat for compactness.

Yes, done that, too! You're right, I was focusing on traditional ITC 
Forths, which I think is what Rod was talking about. In the 70's and 
80's we thought ITC was about the best you could do with a conventional 
implementation, and that turned out not to be the case.

Cheers,
Elizabeth

-- 
==================================================
Elizabeth D. Rather   (US & Canada)   800-55-FORTH
FORTH Inc.                         +1 310.999.6784
5959 West Century Blvd. Suite 700
Los Angeles, CA 90045
http://www.forth.com

"Forth-based products and Services for real-time
applications since 1973."
==================================================

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

#26641

On Fri, 18 Oct 2013 10:01:00 -0400, Stefan Mauerhofer  
<smauerhofer@androsoft.ch> wrote:

>
>> But, be forewarned that most guys either code an OS in
>> x86 assembly or C.  Very few use C++.  Very few develop
>> for non-x86.  I've never heard of Forth being used, probably
>> because Forth either is an OS already or isn't, depending
>> on the implementation.
>
> There is no C required to run a machine. I've implemented a
> 64 bit (native Forth) operating system for the PC using
> assembler, Forth-like macros and Forth only.

In general, I already said that.

There are some people who implement the OS portion of Forth
for their version of Forth.  One of many such examples:

Andy Valencia's ForthOS
http://www.forthos.org/

He recently posted to alt.os.development, BTW.


Rod Pemberton

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

Page 2 of 3 — ← Prev page 1 [2] 3  Next page →

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

csiph-web