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

Groups > comp.lang.forth > #29203 > unrolled thread

Maximum heap memory available to GForth when allot-ing?

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

Contents

  Maximum heap memory available to GForth when allot-ing? Assad Ebrahim <assad.ebrahim@alum.swarthmore.edu> - 2014-04-01 00:10 +0100
    Re: Maximum heap memory available to GForth when allot-ing? David Kuehling <dvdkhlng@posteo.de> - 2014-04-01 06:34 +0200
      Re: Maximum heap memory available to GForth when allot-ing? Assad Ebrahim <assadebrahim2000@gmail.com> - 2014-04-01 00:19 -0700
    Re: Maximum heap memory available to GForth when allot-ing? anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2014-04-01 15:53 +0000
      Re: Maximum heap memory available to GForth when allot-ing? Assad Ebrahim <assadebrahim2000@gmail.com> - 2014-04-01 13:12 -0700
        Re: Maximum heap memory available to GForth when allot-ing? Bernd Paysan <bernd.paysan@gmx.de> - 2014-04-01 23:06 +0200
          Re: Maximum heap memory available to GForth when allot-ing? Assad Ebrahim <assadebrahim2000@gmail.com> - 2014-04-01 14:42 -0700
          Re: Maximum heap memory available to GForth when allot-ing? Assad Ebrahim <assadebrahim2000@gmail.com> - 2014-04-01 18:04 -0700
            Re: Maximum heap memory available to GForth when allot-ing? David Kuehling <dvdkhlng@posteo.de> - 2014-04-02 09:41 +0200
              Re: Maximum heap memory available to GForth when allot-ing? Bernd Paysan <bernd.paysan@gmx.de> - 2014-04-02 12:27 +0200
                Re: Maximum heap memory available to GForth when allot-ing? Assad Ebrahim <assadebrahim2000@gmail.com> - 2014-04-02 10:35 -0700
                  Re: Maximum heap memory available to GForth when allot-ing? Bernd Paysan <bernd.paysan@gmx.de> - 2014-04-02 20:49 +0200
                    Re: Maximum heap memory available to GForth when allot-ing? Assad Ebrahim <assadebrahim2000@gmail.com> - 2014-04-02 16:44 -0700
                      Re: Maximum heap memory available to GForth when allot-ing? Andrew Haley <andrew29@littlepinkcloud.invalid> - 2014-04-03 05:12 -0500
                        Re: Maximum heap memory available to GForth when allot-ing? Mark Wills <markwills1970@gmail.com> - 2014-04-03 04:14 -0700
                          Re: Maximum heap memory available to GForth when allot-ing? Andrew Haley <andrew29@littlepinkcloud.invalid> - 2014-04-03 08:26 -0500
                          Re: Maximum heap memory available to GForth when allot-ing? Bernd Paysan <bernd.paysan@gmx.de> - 2014-04-03 23:04 +0200
                            Re: Maximum heap memory available to GForth when allot-ing? "Rod Pemberton" <dont_use_email@xnothavet.cqm> - 2014-04-04 02:36 -0400
                              Re: Maximum heap memory available to GForth when allot-ing? Andrew Haley <andrew29@littlepinkcloud.invalid> - 2014-04-04 03:38 -0500
                                Re: Maximum heap memory available to GForth when allot-ing? Bernd Paysan <bernd.paysan@gmx.de> - 2014-04-04 15:19 +0200
                                Re: Maximum heap memory available to GForth when allot-ing? "Rod Pemberton" <dont_use_email@xnothavet.cqm> - 2014-04-04 18:03 -0400
                                  Re: Maximum heap memory available to GForth when allot-ing? Andrew Haley <andrew29@littlepinkcloud.invalid> - 2014-04-05 09:12 -0500
                                    Re: Maximum heap memory available to GForth when allot-ing? Assad Ebrahim <assad.ebrahim@alum.swarthmore.edu> - 2014-04-05 15:37 +0100
                          Re: Maximum heap memory available to GForth when allot-ing? hughaguilar96@yahoo.com - 2014-04-07 23:25 -0700
                            Re: Maximum heap memory available to GForth when allot-ing? "Rod Pemberton" <dont_use_email@xnothavet.cqm> - 2014-04-08 06:13 -0400
                        Re: Maximum heap memory available to GForth when allot-ing? Assad Ebrahim <assad.ebrahim@alum.swarthmore.edu> - 2014-04-03 12:39 +0100
                          Re: Maximum heap memory available to GForth when allot-ing? Andrew Haley <andrew29@littlepinkcloud.invalid> - 2014-04-03 08:33 -0500
                      Re: Maximum heap memory available to GForth when allot-ing? albert@spenarnc.xs4all.nl (Albert van der Horst) - 2014-04-03 10:47 +0000
    Re: Maximum heap memory available to GForth when allot-ing? hughaguilar96@yahoo.com - 2014-04-01 16:36 -0700

Page 1 of 2  [1] 2  Next page →

#29203 — Maximum heap memory available to GForth when allot-ing?

Hello -- 


If I run GForth in default mode, i.e. with no command line switches, I
cannot allot more than about 500k cells (2MB) total before getting a
Dictionary overflow error.

This is GForth (v0.7) running in a console window on Windows XP
32-bits...

However, if I use the startup switch   -m size   I can set the
dictionary size to be however large I want, which allows me to allot
more.

So at a practical level, this switch allows me to proceed with my work
just fine.  

But I am wondering why I even have to adjust the dictionary size at
all.

I would have thought that using allot would put a request to the OS to
return free memory, which could be from RAM or from Virtual Memory.  

And given that my computer has 2GB RAM and in excess of 10GB free disk
space available to Virtual Memory, I was expecting that I could allot
much more than 500k cells without needing to adjust the dictionary
space.

What am I missing?

Does the command create w N cells allot actually reserve N cells in
the dictionary?

Thanks in advance!
Assad

[toc] | [next] | [standalone]

#29207

>>>>> "Assad" == Assad Ebrahim <assad.ebrahim@alum.swarthmore.edu> writes:

> But I am wondering why I even have to adjust the dictionary size at
> all.

> I would have thought that using allot would put a request to the OS to
> return free memory, which could be from RAM or from Virtual Memory.

> And given that my computer has 2GB RAM and in excess of 10GB free disk
> space available to Virtual Memory, I was expecting that I could allot
> much more than 500k cells without needing to adjust the dictionary
> space.

> What am I missing?

The region accessed via ALLOT is guaranteed to be contiguous (the ANS
standard defines it that way) and behaves like a large stack (that is
growing towards higher addresses).  I.e. you can repeatedly call ALLOT
to increase memory of one arrays, for example, and free memory allocated
with ALLOT by just running ALLOT on the negative number of bytes
previously ALLOTed.

If you don't need the guarantee of contiguity, use ALLOCATE [1].  This
directly requests memory from the OS and won't depend on the "-m"
command line switch.  You free the memory allocated via ALLOCATE with
FREE.

> Does the command create w N cells allot actually reserve N cells in
> the dictionary?

Yes it does.  

cheers,

David

[1] http://www.complang.tuwien.ac.at/forth/gforth/Docs-html/Heap-Allocation.html#index-allocate-_0040var_007b-u-_002d_002d-a_002daddr-wior--_007d--memory-823

-- 
GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg
Fingerprint: B63B 6AF2 4EEB F033 46F7  7F1D 935E 6F08 E457 205F

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

#29213

On Tuesday, April 1, 2014 5:34:54 AM UTC+1, David Kuehling wrote:
> >>>>> Assad  writes:
> 
> > But I am wondering why I even have to adjust the dictionary size at
> > all.
> 
> > I would have thought that using allot would put a request to the OS to
> > return free memory, which could be from RAM or from Virtual Memory.
> 
> > And given that my computer has 2GB RAM and in excess of 10GB free disk
> > space available to Virtual Memory, I was expecting that I could allot
> > much more than 500k cells without needing to adjust the dictionary
> > space.
> 
> > What am I missing?
> 
> The region accessed via ALLOT is guaranteed to be contiguous (the ANS
> standard defines it that way) and behaves like a large stack (that is
> growing towards higher addresses).  I.e. you can repeatedly call ALLOT
> to increase memory of one arrays, for example, and free memory allocated
> with ALLOT by just running ALLOT on the negative number of bytes
> previously ALLOTed.
> 
> If you don't need the guarantee of contiguity, use ALLOCATE [1].  This
> directly requests memory from the OS and won't depend on the "-m"
> command line switch.  You free the memory allocated via ALLOCATE with
> FREE.
> 
> > Does the command create w N cells allot actually reserve N cells in
> > the dictionary?
> 
> Yes it does.  
> 

Thanks David -- that answers it.  

Cheers,
Assad

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

#29229

Assad Ebrahim <assad.ebrahim@alum.swarthmore.edu> writes:
>Hello -- 
>
>
>If I run GForth in default mode, i.e. with no command line switches, I
>cannot allot more than about 500k cells (2MB) total before getting a
>Dictionary overflow error.

You can see with "UNUSED ." how much dictionary can be ALLOTed.  The
default total dictionary size is 1M cells in 0.7.0, the unused area
is a little smaller.

>But I am wondering why I even have to adjust the dictionary size at
>all.
>
>I would have thought that using allot would put a request to the OS to
>return free memory, which could be from RAM or from Virtual Memory.  

As someone else explained, the dictionary is contiguous, so we
allocate it at once.  One some OSs we can ask for extending a memory
area, but even there it is not guaranteed to work.  Also, for UNUSED
to work, we need to know the maximum size of the dictionary.

The default size is relatively small, because Gforth can be installed
on all kinds of systems, some with small memory.  Also, especially on
32-bit systems, if we reserve a lot of address space for the
dictionary, this reduces the amount of address space we can use for
ALLOCATE and other purposes.

>Does the command create w N cells allot actually reserve N cells in
>the dictionary?

Yes, for some meaning of "reserve".

- 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 2013: http://www.euroforth.org/ef13/

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

#29243

On Tuesday, April 1, 2014 4:53:57 PM UTC+1, Anton Ertl wrote:
> Assad Ebrahim writes:
> 
> >I would have thought that using allot would put a request to the OS to
> >return free memory, which could be from RAM or from Virtual Memory.  
> 
> As someone else explained, the dictionary is contiguous, so we
> allocate it at once.  One some OSs we can ask for extending a memory
> area, but even there it is not guaranteed to work.  Also, for UNUSED
> to work, we need to know the maximum size of the dictionary.
> 
> The default size is relatively small, because Gforth can be installed
> on all kinds of systems, some with small memory.  Also, especially on
> 32-bit systems, if we reserve a lot of address space for the
> dictionary, this reduces the amount of address space we can use for
> ALLOCATE and other purposes.
> 

Yes, that makes sense.

> 
> 
> - anton
> 


Anton ---

while you're here -- do you know whether 
a Windows binary exists for GForth 0.7.2?

David K mentioned the abi-code feature that 
seems like it would make x86 assembly coding 
within GForth easier.  

Unfortunately, 0.7.2 appears to have source only...

Thanks in advance,
- Assad

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

#29244

Assad Ebrahim wrote:
> while you're here -- do you know whether
> a Windows binary exists for GForth 0.7.2?

The reason for 0.7.1 and 0.7.2 were problems with more recent GCCs and gdb 
(used on amd64 for disassembly).  Therefore, I didn't create a new Windows 
binary.  There is no user-visible change between 0.7.0 and 0.7.2 on Windows.

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

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

#29245

On Tuesday, April 1, 2014 10:06:36 PM UTC+1, Bernd Paysan wrote:
> Assad Ebrahim wrote:
> 
> > while you're here -- do you know whether
> > a Windows binary exists for GForth 0.7.2?
> 
> The reason for 0.7.1 and 0.7.2 were problems with more recent GCCs and gdb 
> (used on amd64 for disassembly).  Therefore, I didn't create a new Windows 
> binary.  There is no user-visible change between 0.7.0 and 0.7.2 on Windows.
> 
> -- 
> Bernd Paysan


I see.  

The feature I'm after is related to abi-code mentioned 
here by David K:

https://groups.google.com/d/msg/comp.lang.forth/ib82TjPl_c8/TIcsyxVxufQJ

I've been using GForth 0.7.0.

To try to be sure I've not overlooked it in some vocabulary, 
I did a grep search for abi-code through the gforth\ directory 
but didn't find any matches.

Hence the assumption that it must be in a version later than 0.7.0?

- Assad

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

#29255

On Tuesday, April 1, 2014 10:06:36 PM UTC+1, Bernd Paysan wrote:
> 
> The reason for 0.7.1 and 0.7.2 were problems with more recent GCCs and gdb 
> (used on amd64 for disassembly).  Therefore, I didn't create a new Windows 
> binary.  There is no user-visible change between 0.7.0 and 0.7.2 on Windows.
> 

Fair enough.

So I've tracked down the version that has abi-code: 
it's GForth 0.7.9.  
( http://downloads.qi-hardware.com/people/dvdkhlng/gforth.pdf )

What's the plan for 0.7.9?  

The abi-code stuff seemed fairly complete (at least conceptually)
four years ago based on this thread:
https://groups.google.com/d/msg/comp.lang.forth/BnVnZ1SQoTc/duOn3N8pzckJ

Is it awaiting additional features before it becomes an 0.8.0?

Any chance of an interim Windows binary for it? :D

- Assad

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

#29260

>>>>> "Assad" == Assad Ebrahim <assadebrahim2000@gmail.com> writes:

> Fair enough.

> So I've tracked down the version that has abi-code: it's GForth 0.7.9.
> ( http://downloads.qi-hardware.com/people/dvdkhlng/gforth.pdf )

Ah, no, this is just a pdf generated from a (pretty old?) CVS-Snapshot
of the Gforth sources.  (Note: since then development switched to git).

> What's the plan for 0.7.9?

Good question.  

> The abi-code stuff seemed fairly complete (at least conceptually) four
> years ago based on this thread:
> https://groups.google.com/d/msg/comp.lang.forth/BnVnZ1SQoTc/duOn3N8pzckJ

> Is it awaiting additional features before it becomes an 0.8.0?

I'm also waiting for it.  But just using Git HEAD or a snapshot is not
as bad as it sounds.  I'm almost exclusively using the development
version myself.  Note that there was recently an ABI-CODE regression, so
if you go with Git, go with the latest version.

> Any chance of an interim Windows binary for it? :D

You could install Cygwin and build your version from source.  I almost
don't use windows, I hope there are more qualified people around to help
with or provide a build (try asking on the Gforth mailing list).

cheers,

David
-- 
GnuPG public key: http://dvdkhlng.users.sourceforge.net/dk2.gpg
Fingerprint: B63B 6AF2 4EEB F033 46F7  7F1D 935E 6F08 E457 205F

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

#29262

David Kuehling wrote:

>>>>>> "Assad" == Assad Ebrahim <assadebrahim2000@gmail.com> writes:
> 
>> Fair enough.
> 
>> So I've tracked down the version that has abi-code: it's GForth 0.7.9.
>> ( http://downloads.qi-hardware.com/people/dvdkhlng/gforth.pdf )
> 
> Ah, no, this is just a pdf generated from a (pretty old?) CVS-Snapshot
> of the Gforth sources.  (Note: since then development switched to git).
> 
>> What's the plan for 0.7.9?
> 
> Good question.

Get the more serious changes (new header structure) stabilized enough to 
roll out a release.  That of course requires testing.  Testing in free 
software means volunteers who use it.  So if you want to make it happen, 
stop using 0.7.0, start using Git.

>> The abi-code stuff seemed fairly complete (at least conceptually) four
>> years ago based on this thread:
>> https://groups.google.com/d/msg/comp.lang.forth/BnVnZ1SQoTc/duOn3N8pzckJ
> 
>> Is it awaiting additional features before it becomes an 0.8.0?
> 
> I'm also waiting for it.  But just using Git HEAD or a snapshot is not
> as bad as it sounds.  I'm almost exclusively using the development
> version myself.  Note that there was recently an ABI-CODE regression, so
> if you go with Git, go with the latest version.

Yes, the ABI-CODE stuff was an example of what needed to be fixed before 
making a distribution.  We had a speed regression (thanks to recent GCCs), 
and the Android version, which is a highlight of the upcoming 0.8, also 
needed some serious development.  We still have some troubles with the swig-
based library stuff (and the integration of the swig module into swig 
upstream isn't completed).  I've ironed out a number of problems with 
pthread-based multitasking in the past months.

When you have big plans, it takes longer to get a release.  But the git 
version can be used, and I do upload snapshots, which build without having 
to install another (older) version of Gforth.

>> Any chance of an interim Windows binary for it? :D
> 
> You could install Cygwin and build your version from source.  I almost
> don't use windows, I hope there are more qualified people around to help
> with or provide a build (try asking on the Gforth mailing list).

Noone in the Gforth core team actively uses Windows for anything else than 
gaming.  Since last year, the German tax forms can be run on Linux, so I 
don't have any reasons left to boot Windows.

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

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

#29265

On Wednesday, April 2, 2014 11:27:11 AM UTC+1, Bernd Paysan wrote:
> David Kuehling wrote:
> >>>>>> "Assad" == Assad Ebrahim writes:
>
> >> What's the plan for 0.7.9?
> 
> > Good question.
>
> Get the more serious changes (new header structure) stabilized enough to 
> roll out a release.  That of course requires testing.  Testing in free 
> software means volunteers who use it.  So if you want to make it happen, 
> stop using 0.7.0, start using Git.
> 
> 
> >> The abi-code stuff seemed fairly complete (at least conceptually) four
> >> years ago based on this thread:
> >> https://groups.google.com/d/msg/comp.lang.forth/BnVnZ1SQoTc/duOn3N8pzckJ
> 
> >> Is it awaiting additional features before it becomes an 0.8.0?
> 
> > I'm also waiting for it.  But just using Git HEAD or a snapshot is not
> > as bad as it sounds.  I'm almost exclusively using the development
> > version myself.  Note that there was recently an ABI-CODE regression, so
> > if you go with Git, go with the latest version.
> 
> Yes, the ABI-CODE stuff was an example of what needed to be fixed before 
> making a distribution.  
> 
> When you have big plans, it takes longer to get a release.  But the git 
> version can be used, and I do upload snapshots, which build without having 
> to install another (older) version of Gforth.
> 
> >> Any chance of an interim Windows binary for it? :D
> 
> > You could install Cygwin and build your version from source.  I almost
> > don't use windows, I hope there are more qualified people around to help
> > with or provide a build (try asking on the Gforth mailing list).
> 
> No one in the Gforth core team actively uses Windows for anything else than 
> gaming.  Since last year, the German tax forms can be run on Linux, so I 
> don't have any reasons left to boot Windows.
> 

Ok - I hear you ;)

So if I'm interpreting you correctly, this means that 
during development, Windows support is not a priority 
since none of the core team use it.  But presumably the 
GForth goal of "Available on many machines/easy to port" 
means that a significant new release would still have 
Windows support.

Is that right?

Yes, I used to Cygwin previously, but over the years 
brought over to the Windows platform all the tools that 
I valued and used most often -- LaTeX, R, octave, 
programming languages, command line tools, various 
open source stuff, ffmpeg, etc.  

So now I have all of these plumbed in productively in Windows.  

Since most of my recent professional work is done in 
industrial / commercial environments where Linux is 
pretty much non-existent, Cygwin has become less and 
less necessary, and after a while, I just stopped 
maintaining its presence on the platform. 

However, it's probably time to dust off the abandoned 
laptop and consider installing a Linux distro of some 
sort onto it as the primary OS...

- Assad

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

#29267

Assad Ebrahim wrote:

> On Wednesday, April 2, 2014 11:27:11 AM UTC+1, Bernd Paysan wrote:
>> No one in the Gforth core team actively uses Windows for anything else
>> than
>> gaming.  Since last year, the German tax forms can be run on Linux, so I
>> don't have any reasons left to boot Windows.
>> 
> 
> Ok - I hear you ;)
> 
> So if I'm interpreting you correctly, this means that
> during development, Windows support is not a priority
> since none of the core team use it.  But presumably the
> GForth goal of "Available on many machines/easy to port"
> means that a significant new release would still have
> Windows support.
> 
> Is that right?

Yes.

> Yes, I used to Cygwin previously, but over the years
> brought over to the Windows platform all the tools that
> I valued and used most often -- LaTeX, R, octave,
> programming languages, command line tools, various
> open source stuff, ffmpeg, etc.
> 
> So now I have all of these plumbed in productively in Windows.

Shows that RMS was right saying that too good support of Windows is counter-
productive: People will use free software, but stay with the proprietary 
ecosystem.

> Since most of my recent professional work is done in
> industrial / commercial environments where Linux is
> pretty much non-existent, Cygwin has become less and
> less necessary, and after a while, I just stopped
> maintaining its presence on the platform.
> 
> However, it's probably time to dust off the abandoned
> laptop and consider installing a Linux distro of some
> sort onto it as the primary OS...

Maybe.  It's way more fun with Linux on a recent machine with SSD than on an 
old laptop.

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

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

#29268

On Wednesday, April 2, 2014 7:49:19 PM UTC+1, Bernd Paysan wrote:
> Assad Ebrahim wrote:
> > On Wednesday, April 2, 2014 11:27:11 AM UTC+1, Bernd Paysan wrote:
> 
> > Yes, I used to Cygwin previously, but over the years
> > brought over to the Windows platform all the tools that
> > I valued and used most often -- LaTeX, R, octave,
> > programming languages, command line tools, various
> > open source stuff, ffmpeg, etc.
> 
> > So now I have all of these plumbed in productively in Windows.
> 
> Shows that RMS was right saying that too good support of Windows is counter-
> productive: People will use free software, but stay with the proprietary 
> ecosystem.
> 

I think that depends on the goal.  If every FOSS 
software developer's goal was the death of 
all proprietary software everywhere, then yes.

But that would be a zero-sum world and somehow 
I find it hard to see the world as necessarily zero-sum.

Seems to me that the existence of free open-source software
liberates those who wish for that freedom --- and who are 
willing to invest time & effort in themselves & their toolsets.

But that's not for everyone...

So don't know --- is the inner most desire of most FOSS
authors the end of commercial software?  or that 
like-minded people really like what they've done and join
in the journey of discovery that they're on?

But that's a great quote and provokes some thought in 
any event.

- Assad

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

#29269

Assad Ebrahim <assadebrahim2000@gmail.com> wrote:
> On Wednesday, April 2, 2014 7:49:19 PM UTC+1, Bernd Paysan wrote:
>> Assad Ebrahim wrote:
>> > On Wednesday, April 2, 2014 11:27:11 AM UTC+1, Bernd Paysan wrote:
>> 
>> > Yes, I used to Cygwin previously, but over the years
>> > brought over to the Windows platform all the tools that
>> > I valued and used most often -- LaTeX, R, octave,
>> > programming languages, command line tools, various
>> > open source stuff, ffmpeg, etc.
>> 
>> > So now I have all of these plumbed in productively in Windows.
>> 
>> Shows that RMS was right saying that too good support of Windows is
>> counter-productive: People will use free software, but stay with
>> the proprietary ecosystem.
>> 
> 
> I think that depends on the goal.  If every FOSS software
> developer's goal was the death of all proprietary software
> everywhere, then yes.
> 
> But that would be a zero-sum world and somehow I find it hard to see
> the world as necessarily zero-sum.

How would it be zero sum if all software were to be free?

I think Bernd has somewhat mischaracterized Stallman's argument.  IIRC
Stallman said that supporting free software on Windows didn't help
free software because Widows users tend not to contribute.  I don't
think that is universally true, but IME it is a tendency.

> Seems to me that the existence of free open-source software
> liberates those who wish for that freedom --- and who are willing to
> invest time & effort in themselves & their toolsets.
> 
> But that's not for everyone...
> 
> So don't know --- is the inner most desire of most FOSS authors the
> end of commercial software?  or that like-minded people really like
> what they've done and join in the journey of discovery that they're
> on?

Let's first get our facts straight.  I write free software.  It is
also commercial software.  There is no intrinsic conflict between
freedom and commerce.

Stallman's argument is based on ethics: in his view, proprietary
software is immoral.  So yes, of course he wants the end of
proprietary sofware.  As he put it,

"I'm often asked to describe the 'advantages' of free software. But
the word 'advantages' is too weak when it comes to freedom. Life
without freedom is oppression, and that applies to computing as well
as every other activity in our lives. We must refuse to give the
owners of the programs or computing services control over the
computing we do."

Andrew.

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

#29271

On Thursday, 3 April 2014 11:12:30 UTC+1, Andrew Haley  wrote:
> 
> How would it be zero sum if all software were to be free?
> 
Because (the argument goes) it's an economic race to the bottom. 
> 
> 
> I think Bernd has somewhat mischaracterized Stallman's argument.  IIRC 
> Stallman said that supporting free software on Windows didn't help
> free software because Widows users tend not to contribute.  I don't
> think that is universally true, but IME it is a tendency.

Speaking as a 20+ year (pretty much) exclusive Windows user, I would agree with that. Windows users are, in the main, takers. 
> 
 
> 
> Let's first get our facts straight.  I write free software.  It is
> also commercial software.  There is no intrinsic conflict between
> freedom and commerce.

Tell that to Microsoft. Or Oracle. Even better, tell that to Sun. If you can find anyone, that is.
> 
> Stallman's argument is based on ethics: in his view, proprietary
> software is immoral.  So yes, of course he wants the end of
> proprietary sofware.  As he put it,
> "I'm often asked to describe the 'advantages' of free software. But
> the word 'advantages' is too weak when it comes to freedom. Life
> without freedom is oppression, and that applies to computing as well
> as every other activity in our lives. We must refuse to give the
> owners of the programs or computing services control over the
> computing we do."
> 
Oh what a load of leftist bollocks. "Freedom". FFS! Like someone is putting a gun to one's head, forcing people to buy a particular software product. You are free to NOT buy whatever the hell you like. 

"Life without freedom is oppression, and that applies to computing as well as every other activity in our lives."

I'm just speechless at the shite that this guy spouts. And people worship him like he's fucking Che Guevara or something. Spare me.

Enough. Now I'm just getting pissed off! We're 15 posts into this thread and it's already descended into a load of waffly bollocks.

I'm out.

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

#29275

Mark Wills <markwills1970@gmail.com> wrote:
> On Thursday, 3 April 2014 11:12:30 UTC+1, Andrew Haley  wrote:
>> 
>> How would it be zero sum if all software were to be free?
>> 
> Because (the argument goes) it's an economic race to the bottom. 

Sheesh.  Quelle nonsense.  If free software were actually worse than
proprietary software this argument might have some merit as an
explanation.

>> Let's first get our facts straight.  I write free software.  It is
>> also commercial software.  There is no intrinsic conflict between
>> freedom and commerce.
> 
> Tell that to Microsoft. Or Oracle. Even better, tell that to Sun. If
> you can find anyone, that is.

Thy are entitled to an opinion.  However, I work a great deal with
Oracle on free software, and they don't seem to see an intrinsic
conflict either.

>> Stallman's argument is based on ethics: in his view, proprietary
>> software is immoral.  So yes, of course he wants the end of
>> proprietary software.  As he put it,
>> "I'm often asked to describe the 'advantages' of free software. But
>> the word 'advantages' is too weak when it comes to freedom. Life
>> without freedom is oppression, and that applies to computing as well
>> as every other activity in our lives. We must refuse to give the
>> owners of the programs or computing services control over the
>> computing we do."
>> 
> Oh what a load of leftist bollocks.

Leftist?  I'm baffled.  I always thought that freedom of action was
something trumpeted by the right.

> "Freedom". FFS! Like someone is putting a gun to one's head, forcing
> people to buy a particular software product. You are free to NOT buy
> whatever the hell you like.

I think you're missing the point: if there were only unfree software,
then the owners of the software could indeed control what they would
allow their users to do.  The best example of this is Apple's "walled
garden".

> "Life without freedom is oppression, and that applies to computing
> as well as every other activity in our lives."

Of course it does, and as more and more of our activities are mediated
through computer this becomes increasingly important.

> I'm just speechless at the shite that this guy spouts. And people
> worship him like he's fucking Che Guevara or something. Spare me.
> 
> Enough. Now I'm just getting pissed off! We're 15 posts into this
> thread and it's already descended into a load of waffly bollocks.
> 
> I'm out.

Your loss.

Andrew.

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

#29277

Mark Wills wrote:

> On Thursday, 3 April 2014 11:12:30 UTC+1, Andrew Haley  wrote:
>> 
>> How would it be zero sum if all software were to be free?
>> 
> Because (the argument goes) it's an economic race to the bottom.

Ouch.  Free market economy theory tells you that competition drives costs 
down.  The marginal costs of software have always been close to zero 
(additional copies are cheap), so software, once developed, is dirt cheap.  
If there is a free market, of course.  The mercantilistic way copyrigth 
works means that there is no free market.  Well, copyleft is a game-changer, 
and it results in a way more free market than before.

The cost of software is the cost of writing the first copy.  Ok, there are 
some costs of deployment, and there is the "works for me" effect, which 
means that it doesn't yet work for everybody else, and you have to fix the 
bugs that show up only on other's computers.

But in the end, the job and the task which needs payment is to create and 
maintain the software, not the copies.  And as long as people want and need 
better software, and as long as those experts who write software are a 
scarce resource, money will flow.  This is a law of free market economy.

>> I think Bernd has somewhat mischaracterized Stallman's argument.  IIRC
>> Stallman said that supporting free software on Windows didn't help
>> free software because Widows users tend not to contribute.  I don't
>> think that is universally true, but IME it is a tendency.
> 
> Speaking as a 20+ year (pretty much) exclusive Windows user, I would agree
> with that. Windows users are, in the main, takers.

The somewhat strange thing is that the feedback I get from Android users is 
better than any other feedback I got so far.  I don't think they are a 
different kind of people than Windows consumers (after all, it's a 80% 
market share OS), it's just that Google provides the tools to give feedback 
with its Play Store.  This feedback path is a valuable infrastructure for 
software developers, and Google makes it easy enough for people to use it.

>> Let's first get our facts straight.  I write free software.  It is
>> also commercial software.  There is no intrinsic conflict between
>> freedom and commerce.
> 
> Tell that to Microsoft. Or Oracle. Even better, tell that to Sun. If you
> can find anyone, that is.

Sun did a lot of free software stuff, and was doing quite well.  Oracle 
bought them, because Oracle wanted to have one of their free software 
product: Java.  It's still free software.  Seems you don't have a conclusive 
argument.

>> Stallman's argument is based on ethics: in his view, proprietary
>> software is immoral.  So yes, of course he wants the end of
>> proprietary sofware.  As he put it,
>> "I'm often asked to describe the 'advantages' of free software. But
>> the word 'advantages' is too weak when it comes to freedom. Life
>> without freedom is oppression, and that applies to computing as well
>> as every other activity in our lives. We must refuse to give the
>> owners of the programs or computing services control over the
>> computing we do."
>> 
> Oh what a load of leftist bollocks. "Freedom".

"Freedom" is leftist?  Now you are deep in bollocks land.  Very deep.  
Godwin's law deep.  Free market economy is not leftist.  Free speech is not 
leftist.  Unless, of course, you are so deep into the right that it appears 
to be.

> FFS! Like someone is
> putting a gun to one's head, forcing people to buy a particular software
> product. You are free to NOT buy whatever the hell you like.

Actually, Stallman was opposing to the law which made software proprietary 
by default (this was a pretty new law back then, before, software was not 
under copyright).  There was no choice.  He was the one who actually gave us 
the choice back.  We can now decide if we buy proprietary software or get 
free software instead.  He hacked the copyright system so that this could 
become possible.

> "Life without freedom is oppression, and that applies to computing as well
> as every other activity in our lives."
> 
> I'm just speechless at the shite that this guy spouts. And people worship
> him like he's fucking Che Guevara or something. Spare me.
> 
> Enough. Now I'm just getting pissed off! We're 15 posts into this thread
> and it's already descended into a load of waffly bollocks.
> 
> I'm out.

Hopefully.  This wasn't a brilliant contribution.

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

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

#29279

On Thu, 03 Apr 2014 17:04:10 -0400, Bernd Paysan <bernd.paysan@gmx.de>  
wrote:

> Free market economy theory tells you that competition drives
> costs down.

Wrong.

Free market economic theory tells us that competition drives
*prices* down.

Prices and costs simply aren't the same thing.  Competition doesn't
change the costs of inputs.  Competition, without collusion, generally
lowers prices for consumers, i.e., reduces net profit for the companies.

> The marginal costs of software have always been close to zero
> (additional copies are cheap), so software, once developed,
> is dirt cheap.

Dirt isn't cheap here, but the metaphor is well understood.

> If there is a free market, of course.  The mercantilistic way
> [copyrights] works means that there is no free market.

Are you thinking of patents? ...

> Well, copyleft is a game-changer, and it results in a way more
> free market than before.

Free?  Free in what sense?  Free from profit?  Yes, I'd agree on that.

"... way more free-from-profit market than before."

Yes, that's known as sarcasm.

It's wholly unspecified as to *what* became "free" in "way more free":
brain cells, hair coming out, laughter, pants, love, beer, spittle, etc.

> The cost of software is the cost of writing the first copy.  Ok, there
> are some costs of deployment, and there is the "works for me" effect,
> which means that it doesn't yet work for everybody else, and you have
> to fix the bugs that show up only on other's computers.
>
> But in the end, the job and the task which needs payment is to create
> and maintain the software, not the copies.  And as long as people want
> and need better software, and as long as those experts who write software
> are a scarce resource, money will flow.  This is a law of free market
> economy.

You claim that "money will flow" because "experts who write software are
a scarce resource."  That's true.  It's even more true if the software is
closed source and so the quantity of available experts are even fewer.
You say it's the "law of the free market economy".  It's true, whether or
not it can be qualified as a "law."  Yet, despite all that, just above,
you claim that copyleft makes the market "way more free ... than before."
Am I to assume that "economy" has been excised from "free market economy?"
You clearly support capitalism, yet claim socialism is better.  Ironic.

>>> Let's first get our facts straight.  I write free software.  It is
>>> also commercial software.  There is no intrinsic conflict between
>>> freedom and commerce.
>>
>> Tell that to Microsoft. Or Oracle. Even better, tell that to Sun. If you
>> can find anyone, that is.
>
> Sun did a lot of free software stuff, and was doing quite well.  Oracle
> bought them, because Oracle wanted to have one of their free software
> product: Java.  It's still free software.  Seems you don't have a
> conclusive argument.
>

What?  Oracle bought Sun for the cash flow from their maintenance contracts
and retained cash.

>> FFS! Like someone is
>> putting a gun to one's head, forcing people to buy a particular software
>> product. You are free to NOT buy whatever the hell you like.
>
> Actually, Stallman was opposing to the law which made software  
> proprietary
> by default (this was a pretty new law back then, before, software was not
> under copyright).  There was no choice.

I'm really wondering what law that was since that's always been the case
for state law in the U.S.  It's also that way with international law.  It
wasn't always that way with federal law.

The problem with your argument here is that if Stallman truly was opposing
copyrights, he'd be supporting the concept of *not* having *any*  
copyrights.
Instead, what Stallman created was a *license* for a copyright holder which
only retains minimal rights for the copyright holder.  This license can  
only
be used with copyrighted works!  There are many ways you can view  
Stallman's
license.  One of those ways is that it convinces a copyright holder to
forfit their wealth to others who haven't earned it and don't deserve it.
In legal circles, this is called "fraud" or "conversion", i.e., a con or  
theft.

Also, I think you're conflating "proprietary" with "copyrighted".

> He was the one who actually gave us the choice back.  We can now decide
> if we buy proprietary software or get free software instead.  He hacked
> the copyright system so that this could become possible.

If you "get free software instead", and therefore _clearly_ need and want
better software, just whom is paying the salaries of the scarce experts
who write that software?  You didn't...  Charity?

Supporters of free software always do this.  They say: "Look it's free!"
Then, they turn around and support capitalism in their next breathe saying
that "money will still flow [somehow]."  I don't get it.  The two are *not*
compatible economically.  You can't *not* pay for something and also expect
experts who expect to be paid to code it for you.  Someone must pay them
to work.  Who?  Why?

> [That] wasn't a brilliant contribution.

"This" means your contribution.  "That" would mean his contribution.

No offense intended, since you're exceptionally easy to offend and drive
into an insane rage, but your contribution wasn't much more brilliant than
his: confusion of patents with copyrights, irrational circular arguments,
misunderstanding of economic theories, etc.  Maybe, you should refrain from
ranking the contributions of others.  It seems your not qualified to do so.


Rod Pemberton

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

#29280

Rod Pemberton <dont_use_email@xnothavet.cqm> wrote:
> 
> Supporters of free software always do this.  They say: "Look it's
> free!"  Then, they turn around and support capitalism in their next
> breathe saying that "money will still flow [somehow]."

There is no incompatibility between commerce and free software.

> I don't get it.  The two are *not* compatible economically.  You
> can't *not* pay for something and also expect experts who expect to
> be paid to code it for you.  Someone must pay them to work.  Who?
> Why?

Because you can pay for free software: it's free as in freedom, not as
in price.  "Software libre" may be a better term.

Andrew.

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

#29281

Andrew Haley wrote:

> Rod Pemberton <dont_use_email@xnothavet.cqm> wrote:
>> 
>> Supporters of free software always do this.  They say: "Look it's
>> free!"  Then, they turn around and support capitalism in their next
>> breathe saying that "money will still flow [somehow]."
> 
> There is no incompatibility between commerce and free software.

And the only incompatibility between capitalism and freedom is slavery.

>> I don't get it.  The two are *not* compatible economically.  You
>> can't *not* pay for something and also expect experts who expect to
>> be paid to code it for you.  Someone must pay them to work.  Who?
>> Why?
> 
> Because you can pay for free software: it's free as in freedom, not as
> in price.  "Software libre" may be a better term.

The problem people like Rod have is that they confuse goods and service.  
Software is a service: You write it, you add features, you fix bugs, that's 
the work you get paid for.  That's a service.  The copy is free, it's not a 
good.  A good is a physical entity, a service is intangible.  It is very 
obvious if you know your economics 101, that software is not a good, because 
software as such is intangible (the copies exist on physical entities, but 
that's mass storage: a good).

Once you understand that software is a service, you understand how to make 
money with copylefted software.  RedHat, where Andrew is employed, makes 
tons of money.  With free software.

Copyrigth is an early 18th century invention, at the dawn of 
industrialization.  It was a failed attempt in breaking the printer's guild 
monopoly.  The attempt was to grant the author the monopoly, not the 
printer.  It was a failed attempt, because the printers still insisted on 
getting a "total buy-out", just like before, where the author sold his 
manuscripts as "service".  In fact, countries like Germany, which didn't 
have a copyright until the mid-19th century, were way more successful in 
breaking the guild's monopoly.  A lot of "pirated" books (as the content 
industry would say) made printing a very competitive environment, and with a 
similar population, Germany had about 10 times the publications (titles) as 
GB - and 100 times as many books (copies).  Obviously, the incentive of 
copyright is not to print many copies of many authors (which all want their 
pay to deliver a manuscript), but to print few copies of few authors at an 
excessive price.

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

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

Page 1 of 2  [1] 2  Next page →

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

csiph-web