Copyleft and embedded Forth

17 views
Skip to first unread message

Brad Eckert

unread,
Jan 28, 2008, 11:20:30 AM1/28/08
to
I'm still trying to get my head around the Free Software concept. If I
buy a Ford and it uses GPL'd code in it's engine controller, Ford is
required to publish the source code to it's engine controller. I'm
guessing that Ford would rather not have the GM and Chevy engineers
poring over their code so they would buy a license to the otherwise
GPL'd code. No problem.

Something like CDDL allows commercial developers to keep one foot in
the Open Source camp since only selected files are copylefted. If you
change them, you have to publish the source code to your changes. Not
unreasonable. If you write some words (in your own file) that link to
words from a CDDL-licensed file, you don't have to publish your code
because it can stay under whatever license you want.

Between GPL and CDDL, which license would better serve the Open Source
Forth community? GPL seems too strong and CDDL too weak.

Brad

Jean-François Michaud

unread,
Jan 28, 2008, 4:37:51 PM1/28/08
to

I prefer the GPL because it completely blows the lid off source code
availability and makes available useful software to everybody to work
on. This is why the Linux community is flourishing so much. By having
fully available source, you ensure that alot more people are going to
be able to do serious work with the code by building upon it while
ensuring that the derived work is also available for further
enhancements (the more eyeballs, the better). There is front work
effort required but I believe the products rewards are great.

Also, from my understanding, making the code open source doesn't
prevent the software from being sold for use. Basically, the imposed
control is not at the code level anymore but at the service level.
My understanding is that if you do work derived from GPL'd software,
then you need to ensure that the source code is available to whoever
pulls your software. Period. You can't enforce further restrictions on
code availablity but the license doesn't restrict you from charging
for what you offer.

For example, maintenance fees can be charged when selling GPL'd
software to ensure a form of return on the product going out the door.
Software that hasn't been acquired directly by the source of the
product can simply see their request for support disregarded.

This is roughly how linux servers roll. The software is roughly free
but comes bundled with a service agreement that costs you. Companies
specialize in bundling open source software and offering services
associated with maintaining the bundle functional. Or you can take the
risk of rolling your software and it failing on you without anybody to
turn to if it all goes to hell; but then the particular bundle that
you're looking at might not be made available to you (you'll have to
put it together yourself, for example). We turn a product economy into
a service economy. The door is seemlingly open and anybody can
potentially end up being able to acquire anything, even without paying
for it, but in reality, this promotes the propagation of the used
software and it's popularity (Microsoft vs software piracy is a good
example, or in the open source world, Linux being more popular than
Apple (Apple which has much more stringent blocks on Piracy than
Microsoft, in good part due to it's hardware).

1- Proprietary (Microsoft lets hacking continue for a long while)
Result: Microsoft software is exceedingly widespread although
annoyingly controlling due to it's proprietary essence.

2- Proprietary (Apple doesn't let hacking take place)
Result: Less popular simply because less openly exposed to the
public in both code being available and Piracy.

3- Open source (Linux lets the software flow)
Result: Same as Microsoft, the software is becoming exceedingly
widespread and is becomes extremly robust because of the massive
amount of eyeballs that can take a look. We see the software being
more popular that the Apple alternative, even though Mac OS X is more
polished from a user standpoint.

You might also want to consider the LGPL (Lesser GPL) as an
alternative if you want to be more permissive. I prefer not to because
the GPL forces the creation of useful building blocks that the whole
world can use, potentially free of charge. This is conducive to very
productive and creative work. Everybody can work together while
constantly enhancing products, and this, while still making the
economy roll.

Regards
Jean-Francois Michaud

Bruce McFarling

unread,
Jan 28, 2008, 4:54:45 PM1/28/08
to

A two level strategy would entail ...

... aggressively Open Source code, under the GPL, which thereby can
tap any code or library under the GPL.

... permissively Open Source code, under something that provides the
functional equivalent of the LGPL in a Forth context, which thereby
can propagate Open Source code amongst users not willing to release
the result under the GPL ... but, of course, may require licensing
fees to be paid to get the commercial equivalent of what the
aggressively Open Source code can freely use by being released under
the GPL.

http://www.gnu.org/licenses/why-not-lgpl.html

... is the FSF rallying cry to release as many libraries under GPL as
possible to best advantage free software relative to commercial
software.

Rallying cries are great and all, however last I recall it being
discussed, the LGPL is not really a Library GPL or even that much of a
Lesser-hassle GPL for a classic forth under the LGPL and then extended
into the application. However, the LGPL might work for a cross-
compiler if the foundation core of the target code base was public
domain and the host Forth was LGPL.

What would really be needed for a straightforward permissive level GPL
for a classic Forth programming paradigm would seem to be an
Embeddable GPL, where a proprietary extension of a EGPL foundation
could be allowed by the copyright holder, provided that the original
EGPL foundation is also made available, other by inclusion of the
original foundation system or a function allowing the the EGLP system
to be returned to its foundation state.

Albert van der Horst

unread,
Jan 28, 2008, 6:08:25 PM1/28/08
to
In article <6d26f0f4-9ddd-4449...@h11g2000prf.googlegroups.com>,
Brad Eckert <nospaa...@tinyboot.com> wrote:
<SNIP>

>
>Between GPL and CDDL, which license would better serve the Open Source
>Forth community? GPL seems too strong and CDDL too weak.

I promote the commercial ;-) use of ciforth by granting extra rights
to otherwise GPL-ed software. You will find it in de pdf documentation
of lina e.a.
Because one can relate very specifically to Forth, this is much
easier than interpret how a certain license would work out
legally with respect to Forth.

>
>Brad

Groetjes Albert

--

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

ian...@gmail.com

unread,
Jan 28, 2008, 7:39:15 PM1/28/08
to

For someone trying to develop real applications in FORTH the
complexities of the GPL and LGPL persuade me very quickly to use
software with a BSD style license or pure commercial software. With
commercial software you have to ensure you have all the rights you
need....maybe you wish to expose the interpreter. One might also have
a reasonable arrangement with the authors of GPL/LGPL'd software to
ensure the correct legal result. I remember a recent thread regarding
gpl/lgpl and pfe (which we have used in the past), I was surprised by
some of the interpretations when one focusses on FORTH and believe
that others may well have the same opinion if they have invested
(very) heavily in their base code and algorithms.

Of course if your goal is to learn, have fun, show that enthusiasts
can do better than commercial outfits, or "invent" another FORTH there
is not too much wrong with GPL/LGPL. It has obvious advantages in
distributed collaborative software generation,

regards

Ian

Charlie Springer

unread,
Jan 29, 2008, 12:45:25 AM1/29/08
to
Is any of this enforcible? If it is free and open-source why would I want to
pay any attention to "requirements" that I make the code available?

-- Charlie Springer

Anton Ertl

unread,
Jan 29, 2008, 4:56:39 AM1/29/08
to
Brad Eckert <nospaa...@tinyboot.com> writes:
>I'm still trying to get my head around the Free Software concept. If I
>buy a Ford and it uses GPL'd code in it's engine controller, Ford is
>required to publish the source code to it's engine controller. I'm
>guessing that Ford would rather not have the GM and Chevy engineers
>poring over their code

Maybe, maybe not. It seems that lots of companies have found it
profitable to use GPLed code (e.g., a Linux kernel) in their embedded
products. However, instead of second-guessing Ford's engineers, you
should rather think of what your goals are and which license helps you
to achieve it.

>Between GPL and CDDL, which license would better serve the Open Source
>Forth community?

Depends on what goals you want to achieve. The FSF wants to give the
crucial freedoms to as many users as possible ("Open Source" may have
other goals). Their strategy is to use the GPL for libraries that
have no competition or for which the competition is weak. That gives
a competetive advantage to developers of free software over developers
of proprietary software.

E.g., in your example above, if Ford wants to do proprietary code,
they will have to develop their own library at additional cost and
additional time to market, whereas if GM and Toyota decide to do free
software, they can use the library right away, and also profit from
the software that the other company has to release (that's the Open
Source advantage).

- 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 2008: http://www.complang.tuwien.ac.at/anton/euroforth/ef08.html

Brad Eckert

unread,
Jan 29, 2008, 10:12:29 AM1/29/08
to
On Jan 28, 10:45 pm, Charlie Springer <R...@regnirps.com> wrote:
> Is any of this enforcible? If it is free and open-source why would I want to
> pay any attention to "requirements" that I make the code available?
>
This is a popular approach. The odds of getting busted are slim, so
many companies take their chances. If they get caught, they pay a
settlement and move on. Sarbanes-Oxley ups the ante by providing
federal penalties, however. There's always a chance that in the
future, IP trolls will find this fertile feeding ground and data
mining software will be able to detect GPL components in commercial
software.

Just because someone allows a technology to flourish for years doesn't
mean they wont start their asserting rights a few years down the road.
Nobody heard a peep out of Fraunhofer as the MP3 format proliferated
on the Internet. Then came the "Oh, by the way" notices from the
lawyers.

BTW, if you keep the code in-house and don't distribute it, GPL
doesn't require you to to release the source code.

Brad

Brad Eckert

unread,
Jan 29, 2008, 10:33:49 AM1/29/08
to
On Jan 29, 2:56 am, an...@mips.complang.tuwien.ac.at (Anton Ertl)
wrote:

>
> E.g., in your example above, if Ford wants to do proprietary code,
> they will have to develop their own library at additional cost and
> additional time to market, whereas if GM and Toyota decide to do free
> software, they can use the library right away, and also profit from
> the software that the other company has to release (that's the Open
> Source advantage).
>
I'm still working to overcome 30 years of psychological conditioning
in consumerism, but this makes sense. If a company uses open source,
they could help their competitors gain market share that they wouldn't
otherwise have. OTOH, an increasing market size offsets mitigates this
disadvantage.

Brad

Brad Eckert

unread,
Jan 29, 2008, 11:29:02 AM1/29/08
to
On Jan 28, 9:20 am, Brad Eckert <nospaambr...@tinyboot.com> wrote:
> I'm still trying to get my head around the Free Software concept. If I
> buy a Ford and it uses GPL'd code in it's engine controller, Ford is
> required to publish the source code to it's engine controller.

However, the VHDL/Verilog component of an embedded system isn't
covered by GPL. GPL links to it, not the other way around. So special
sauce can always be encapsulated in hardware.

Brad

Jean-François Michaud

unread,
Jan 29, 2008, 8:42:40 PM1/29/08
to

Right, but there is also a higher level that seems to be mostly
disregarded in this discussion. The service level is quite significant
in the open source world and expertise required to bundle multiple
components together into a seamless package is not insignificant. That
expertise and maintenance proposal can be sold for quite the price. A
company can do nothing more than simply making sure a package is
seamlessly integrated and it can roll. It's relatively useful to think
in terms of service rather than product in the open source context.
Google takes it to the next level. Free service (fee for enhanced
service through priority bumping on searches for example). This would
be the open source pendant of services; everybody benefits from the
enabling service and the company still makes money. This effectively
propels the company and products/service reputation to stellar heights
if the product/service is sufficiently useful simply because more
people can easily be in contact with it. You can start pulling harder
too once the service is omnipresent and when it becomes difficult to
envision life without it...or not ;-).

What would you do if the companies that own search engines decided to
start charging for searching the net? You'd have to pick your favorite
and go with the flow simply because it's difficult to envision doing
anything without a search engine.

Regards
Jean-Francois Michaud

none Byron Jeff

unread,
Jan 30, 2008, 4:50:12 PM1/30/08
to
In article <62ef541b-337a-449c...@h11g2000prf.googlegroups.com>,

Jean-François Michaud <com...@comcast.net> wrote:
>On Jan 28, 8:20 am, Brad Eckert <nospaambr...@tinyboot.com> wrote:
>> I'm still trying to get my head around the Free Software concept. If I
>> buy a Ford and it uses GPL'd code in it's engine controller, Ford is
>> required to publish the source code to it's engine controller.

AFAICT that's correct.

>> I'm
>> guessing that Ford would rather not have the GM and Chevy engineers
>> poring over their code so they would buy a license to the otherwise
>> GPL'd code. No problem.

Probably.

>> Something like CDDL allows commercial developers to keep one foot in
>> the Open Source camp since only selected files are copylefted. If you
>> change them, you have to publish the source code to your changes. Not
>> unreasonable. If you write some words (in your own file) that link to
>> words from a CDDL-licensed file, you don't have to publish your code
>> because it can stay under whatever license you want.

It's taken me a long time (something like 5 years) to understand that
Open Source needs and has different views on code that serves as system
infrastructure and code that serves as application. Infrastructure code
that is a common base that many can use needs a license that serves two
purposes: To ensure that everyone has complete access to the
infrastructure (and its modifictions) and to be neutral in terms of code
licensing for code that uses the infrastructure. OTOH for applications
built on the infrastructure, it's the wild, wild west where everyone is
free to compete as they see fit and use whatever license they see fit.

Let me address the points below, then I'll get back to the issue of
embedded systems...


>I prefer the GPL because it completely blows the lid off source code
>availability and makes available useful software to everybody to work
>on. This is why the Linux community is flourishing so much. By having
>fully available source, you ensure that alot more people are going to
>be able to do serious work with the code by building upon it while
>ensuring that the derived work is also available for further
>enhancements (the more eyeballs, the better). There is front work
>effort required but I believe the products rewards are great.

But I think that's short sighted. Fundamentally none of the
infrastructure systems for Linux are pure GPL. The kernel is GPL but is
has a usage exception. The primary libraries are all LGPL or BSD. X uses
the MIT license IIRC. So applications with all types of licenses from
GPL to commercial can work within the Linux framework.

Linux would be a totally different animal if Linus had insisted that all
programs that use the Linux kernel be GPLed.

>Also, from my understanding, making the code open source doesn't
>prevent the software from being sold for use. Basically, the imposed
>control is not at the code level anymore but at the service level.
>My understanding is that if you do work derived from GPL'd software,
>then you need to ensure that the source code is available to whoever
>pulls your software. Period. You can't enforce further restrictions on
>code availablity but the license doesn't restrict you from charging
>for what you offer.

That's a strawman. When code can be copied freely, its value tends
towards free. It's the same problem that commercial software, music, and
movies are having right now. The only difference with free software is
that the license allows for free copying.

So for commercial developers this is a death knell. They want to retain
inherent value for their software by restricting copying.

>For example, maintenance fees can be charged when selling GPL'd
>software to ensure a form of return on the product going out the door.
>Software that hasn't been acquired directly by the source of the
>product can simply see their request for support disregarded.

Again commercial software developers want to make money by selling
software, not services.

>This is roughly how linux servers roll. The software is roughly free
>but comes bundled with a service agreement that costs you. Companies
>specialize in bundling open source software and offering services
>associated with maintaining the bundle functional. Or you can take the
>risk of rolling your software and it failing on you without anybody to
>turn to if it all goes to hell; but then the particular bundle that
>you're looking at might not be made available to you (you'll have to
>put it together yourself, for example). We turn a product economy into
>a service economy. The door is seemlingly open and anybody can
>potentially end up being able to acquire anything, even without paying
>for it, but in reality, this promotes the propagation of the used
>software and it's popularity (Microsoft vs software piracy is a good
>example, or in the open source world, Linux being more popular than
>Apple (Apple which has much more stringent blocks on Piracy than
>Microsoft, in good part due to it's hardware).

The only hope in the particular arena that we're talking about is the
fact that the software is embedded into a product. So there is inherent
value in the physical hardware.

But trust me, commercial developers do not want to function in a service
economy. They want their dollars up front for the software sale and on
the back end for service/support.

The GPL simply doesn't accomodate their way of thinking. Now they may
end up dying in a free market where the competition is giving their
software away, but personally in years and years of open source advocacy
I've never seen any commercial developer convinced that they can do
better in a free software/service economy.

>1- Proprietary (Microsoft lets hacking continue for a long while)
> Result: Microsoft software is exceedingly widespread although
>annoyingly controlling due to it's proprietary essence.
>
>2- Proprietary (Apple doesn't let hacking take place)
> Result: Less popular simply because less openly exposed to the
>public in both code being available and Piracy.
>
>3- Open source (Linux lets the software flow)
> Result: Same as Microsoft, the software is becoming exceedingly
>widespread and is becomes extremly robust because of the massive
>amount of eyeballs that can take a look. We see the software being
>more popular that the Apple alternative, even though Mac OS X is more
>polished from a user standpoint.
>

The real problem is that you can't prove the above point. By all
statistical and sales measures Microsfot still owns an overwhelming
share of the market, followed by Apple. Linux even after nearly 17 years
of progress still lags far behind in both market and mindshare.

>You might also want to consider the LGPL (Lesser GPL) as an
>alternative if you want to be more permissive.

I'll get to the LGPL in a moment, but it too doesn't work in the
environment that Brad is discussing.

>I prefer not to because
>the GPL forces the creation of useful building blocks that the whole
>world can use, potentially free of charge. This is conducive to very
>productive and creative work. Everybody can work together while
>constantly enhancing products, and this, while still making the
>economy roll.

OK. I've let you say your peace, Jean-Francois. Now I'll finish my
points:

1) Linux isn't a pure GPL system. The infrastructure is a collection of
exceptions and other licenses that permit the mixing of applications of
difference license stripes.

2) One reason that it works is because the LGPL, upon which many
libraries use, allow for the dynamic linking of applications which are
neither GPL or LGPL with them. In short anyone can use the
infrastructure.

3) However by the same token, the libraries (and the kernel BTW) are
protected. Modifications to those entities must have published source
code. The end result is that everyone benefits from the infrastructure
without hindering the free development (licensewise) of applications
that merely use that infrastructure.

4) The problem is that it doesn't work in an embedded systems
environment, of which in some ways a typical Forth system qualifies.
Consider if you build a useful collection of words that provides an
infrastructure for building applications. Each license has an impact
upon both the source for the infrastructure and the source for the
applications using that infrastructure:

GPL: easy. All modifications to the infrastructure must be published. In
addition all source for any applications must also be published.
Everything is open. However, no means of having any other different type
of licensed code is allowed.

LGPL: not as easy. Under normal circumstances. the licensing of the
infrastructure is GPL and the license for the application is
unencumbered. So like the GPL source modifications for the
infrastructure must be published. That's fine. However the LGPL has a
restriction that states that downline users must have the right to
upgrade the LGPL portion of a system. Typically this is done my dynamic
linking and has no impact upon the application. But in an embedded
system (and a typical Forth system) the infrastructure and the
application are comingled. So that means the only effective way to build
a system with an updated LGPL infratsructure is to have the source for
the infrastructure (no problem), AND THE SOURCE FOR THE APPLICATION!
Oops! So now you're back to a GPL situation.

BSD: easy but not good. BSD fundamentally removes all restrictions
except for the removal of the copyright notice. So any developer can do
deep modification of the infrastructure and not share any of it. Same
for the application of course. The problem is that a deep pocketed
developer can take an existing infrastructure, modify it to the point
where it is incompatible with the original infrastructure, then release
the new infrastructure and application with no source. So now no one but
the developer gets benefit of the changed infrastructure. Code forks and
market takeovers can then ensue. Again not good.

CDDL: Frankly it's the embedded systems license that I've been looking
for. Like the GPL and LGPL, the infrastructure source and its
modifications must be public. But applications code has no encumbrance.
In addition the LGPL relink requirement is removed. There is a small
loss for a downstream developer who wants to update the CDDL
infrastructure but doesn't necessarily have access to the applications
code to do so. But since everyone must contribute to the infrastructure,
a strong protected base for everyone to develop upon is built. No code
forks, no swiping of infrastructure turf.

CDDL has balance. It recognizes that now everyone is going to be able to
free their codebases. However it does protect its own codebase. And it
facilitates mixing of codebases that function at different levels of
licensing.

Just my three cents.

BAJ

Jean-François Michaud

unread,
Jan 31, 2008, 2:14:20 AM1/31/08
to
On Jan 30, 1:50 pm, byron@upstairs.(none) (Byron Jeff) wrote:
> In article <62ef541b-337a-449c-a869-f80275a67...@h11g2000prf.googlegroups.com>,

Is it? Of maybe not sufficiently precise. I do understand that
limitations are applied and that Linux is not a full GPL model, but I
get a clear vibe that without the openness of the model on which Linux
is built, that none of the wonderful things we see happening with this
OS would have been possible. Free and open seems to make, in my view,
a clear-cut difference; whether you think a full GPL model can survive
through, for example, a more service oriented model is up for grabs
really, but I think it's safe to contend that the essence of openness
that the GPL enforces triggers the channeling of a very serious
information flow towards it. Because such a heavy flow is created,
robustness and usefulness ensues. Without the GPL, we are talking a
different game completely. Creating such an information flow on a
proprietary application is simply out of the question without a
gargantuan cash-flow.

Fundamentally none of the
> infrastructure systems for Linux are pure GPL. The kernel is GPL but is
> has a usage exception. The primary libraries are all LGPL or BSD. X uses
> the MIT license IIRC. So applications with all types of licenses from
> GPL to commercial can work within the Linux framework.
>
> Linux would be a totally different animal if Linus had insisted that all
> programs that use the Linux kernel be GPLed.

Indeed. It certainly would have been interesting to see.

> >Also, from my understanding, making the code open source doesn't
> >prevent the software from being sold for use. Basically, the imposed
> >control is not at the code level anymore but at the service level.
> >My understanding is that if you do work derived from GPL'd software,
> >then you need to ensure that the source code is available to whoever
> >pulls your software. Period. You can't enforce further restrictions on
> >code availablity but the license doesn't restrict you from charging
> >for what you offer.
>
> That's a strawman. When code can be copied freely, its value tends
> towards free.

This is, from my perspective, conducive to a widespread use/adoption
of the software. The reputation goes up and other products/services
can be sold once the information associated with your product/service
is anchored in many minds. Not everybody will try to acquire stuff for
free whenever possible; you have to place yourself in a different
state of mind that is somewhat incompatible with the current state of
mind and economic model. The current model asks you to pull the
blanket hard on your side (I give you A for the price of B) whereas
the newer model asks you to pull on a different blanket (I give you A
for free and you can return the favor by contributing also in your own
way if you feel it's relevant but If I absolutely want to make money,
I might charge a service fee to stay in business). I can only guess
that you find the contribution box to be a ridiculous concept. I
actually give out money to folks that provide me with what I feel is
useful information when they provide means for me to do so. The
information is available for free for everybody to see, but they don't
strictly enforce being paid for what they do.

It's the same problem that commercial software, music, and
> movies are having right now. The only difference with free software is
> that the license allows for free copying.

Only a problem because they want it to be a problem. I still buy CDs
even though music is available for free online (I'm actually blessed,
or cursed depending on how you see it, with a penchant for higher
quality recordings. I simply can not handle listening to MP3 and I
despise compression (it actually hurts my ears when listened through
my equipment) so I end up buying (hopefully) good quality recordings
of at least CD quality and I rip them to the computer. You can't
usually get full quality recordings for free online because of the
amount of space each file takes (roughly 60 Mb a song on a .wav
format; maybe 30Mb on a lossless format like FLAC). I still rent and
buy movies even if I can get them for free online but I did switch to
a more flexible model that actually allows you to view any amount of
movies online. As long as you can download and have the time, you can
watch. Somehow, even though they offer a lot, though the use of a much
more flexible model that's closer in nature to openness, I'm certain
that Netflix is not going to end up poor.

It's because of this type of control that I completely dropped
Windows. Vista hit the last nail on the head and I will never use
Windows again for personal use. It's Apple from now on or Linux. I'll
live.

> So for commercial developers this is a death knell. They want to retain
> inherent value for their software by restricting copying.

Certainly a remnant of the vestigial view that insists for the need to
receive **immediate** compensation for work. This works well for
physical goods, not so well for products of the mind.

> >For example, maintenance fees can be charged when selling GPL'd
> >software to ensure a form of return on the product going out the door.
> >Software that hasn't been acquired directly by the source of the
> >product can simply see their request for support disregarded.
>
> Again commercial software developers want to make money by selling
> software, not services.

I'm a developer and I love coding, but I would willingly make all of
the code I write open source so that others can benefit from it.
Also, I'm certainly not against the idea of selling services
accompanying the software I would open source. Enhancing the
functionality of open source software can come for a fee; not
everybody that has access to open source software wants to enhance its
functionality by learning to code. Compensation is not immediate and
that clearly offends you and probably many others as well. It does
require a thought shift.

> >This is roughly how linux servers roll. The software is roughly free
> >but comes bundled with a service agreement that costs you. Companies
> >specialize in bundling open source software and offering services
> >associated with maintaining the bundle functional. Or you can take the
> >risk of rolling your software and it failing on you without anybody to
> >turn to if it all goes to hell; but then the particular bundle that
> >you're looking at might not be made available to you (you'll have to
> >put it together yourself, for example). We turn a product economy into
> >a service economy. The door is seemlingly open and anybody can
> >potentially end up being able to acquire anything, even without paying
> >for it, but in reality, this promotes the propagation of the used
> >software and it's popularity (Microsoft vs software piracy is a good
> >example, or in the open source world, Linux being more popular than
> >Apple (Apple which has much more stringent blocks on Piracy than
> >Microsoft, in good part due to it's hardware).
>
> The only hope in the particular arena that we're talking about is the
> fact that the software is embedded into a product. So there is inherent
> value in the physical hardware.
>
> But trust me, commercial developers do not want to function in a service
> economy.

I would.

They want their dollars up front for the software sale and on
> the back end for service/support.
>
> The GPL simply doesn't accomodate their way of thinking.

Interesting, you seem to hold dominion over the thoughts of all
software developers, including my own, which, apparently, from where
I'm standing, are not at all aligned with what you're saying.

Now they may
> end up dying in a free market where the competition is giving their
> software away, but personally in years and years of open source advocacy
> I've never seen any commercial developer convinced that they can do
> better in a free software/service economy.

Do you realize how difficult it is to switch over to a different
thought model, and in particular when everybody around you is
screaming that it won't work; not necessarily because they know
better, but simply because they are absolutely convinced that it can
not function any differently than what they are accustomed to? This
doesn't invalidate the current model, it works, but that certainly
shouldn't imply that the newer and potentially more flexible model
can't function as well. A hybrid model is also certainly useful to
consider; In the meantime, I'll try pushing hard on the free side too
try and facilitate the transition.

> >1- Proprietary (Microsoft lets hacking continue for a long while)
> > Result: Microsoft software is exceedingly widespread although
> >annoyingly controlling due to it's proprietary essence.
>
> >2- Proprietary (Apple doesn't let hacking take place)
> > Result: Less popular simply because less openly exposed to the
> >public in both code being available and Piracy.
>
> >3- Open source (Linux lets the software flow)
> > Result: Same as Microsoft, the software is becoming exceedingly
> >widespread and is becomes extremly robust because of the massive
> >amount of eyeballs that can take a look. We see the software being
> >more popular that the Apple alternative, even though Mac OS X is more
> >polished from a user standpoint.
>
> The real problem is that you can't prove the above point. By all
> statistical and sales measures Microsfot still owns an overwhelming
> share of the market, followed by Apple. Linux even after nearly 17 years
> of progress still lags far behind in both market and mindshare.

Linux is not a company, so it's certainly difficult to get a feel for
it's popularity by looking at market shares, but the mind boggling
diversity that we see in the Linux world seems to me to be a good
testament to it's popularity and general well being. It's certainly
easier to find information on how to get things done in Linux than it
is for OSX. There are many things to consider here but it's certainly
an interesting indicator.

> market takeovers can then...
>
[SNIP]

I'm not particularly savvy of the types of law related limitations
people like to impose on embedded systems or embedded system software.
I'm simply advocating a freer (in the freedom sense, not in the free
from fee sense, which you seem to mix up), more open, model. Is it
entirely sufficient in itself, maybe not, but I certainly think it's a
push in a more appropriate direction.

Regards
Jean-Francois Michaud

Albert van der Horst

unread,
Jan 31, 2008, 5:09:01 AM1/31/08
to
In article <_q6dnfqOkpKJbD3a...@comcast.com>,
none) (Byron Jeff <byron@upstairs.> wrote:
<SNIP>

>
>Again commercial software developers want to make money by selling
>software, not services.

This is a flawed business model that in the longer run can only
be imposed by a military dictatorship.
(But we're getting there ;-) ).

<SNIP>


>
>1) Linux isn't a pure GPL system. The infrastructure is a collection of
>exceptions and other licenses that permit the mixing of applications of
>difference license stripes.

It is called GNU/linux for a reason. All the tools, in particular the
compiler, were there, and in the end amount to more effort than the OS.
In other words if there was not a Finnish student, a Korean student would
have created an OS on that base. This is the essence.

Then there were some corners cut to make things practical.

<SNIP>

>
>BAJ

Stephen Pelc

unread,
Jan 31, 2008, 6:54:38 AM1/31/08
to
On Wed, 30 Jan 2008 15:50:12 -0600, byron@upstairs.(none) (Byron Jeff)
wrote:

>But trust me, commercial developers do not want to function in a service
>economy. They want their dollars up front for the software sale and on
>the back end for service/support.

There are several examples (including at least one Forth application)
where the reverse is true. The big advantage of a service economy,
especially a price/month model, is that you get your development
money when you need it. You get a revenue stream *while* you are
doing development, not *afterwards*. From a toolmaker's view,
the little and often release model it encourages means that you
have fewer clients locked into a ten-year old system that requires
a big investment to move away from.

It's a very different model, and needs a very different set of
relationships. When it is appropriate, it works really well.
MPE likes them, but they are not for everyone.

Stephen

--
Stephen Pelc, steph...@mpeforth.com
MicroProcessor Engineering Ltd - More Real, Less Time
133 Hill Lane, Southampton SO15 5AF, England
tel: +44 (0)23 8063 1441, fax: +44 (0)23 8033 9691
web: http://www.mpeforth.com - free VFX Forth downloads

Brad Eckert

unread,
Jan 31, 2008, 10:35:28 AM1/31/08
to
>
> Certainly a remnant of the vestigial view that insists for the need to
> receive **immediate** compensation for work. This works well for
> physical goods, not so well for products of the mind.
>
Do you perceive that the public mind is coming around to the idea that
software is a service. If so, the open software model might have a
chance. For example, do you leave a restaurant without tipping the
waitress or a resort without tipping the concierge? Most people don't
because it's not customary and considered rude. Open source sites
might do well to have an icon of a concierge in a tux as well as a
"tips" jar you can click to leave a tip.

Brad

Jean-François Michaud

unread,
Jan 31, 2008, 11:30:05 AM1/31/08
to

Right on the money! ;-). That would be the general sense I'm getting,
but this conceptual shift is difficult to make for many and seems to
have left a mark the size of a 2x4 in the face of our friend NBJ.

Regards
Jean-Francois Michaud

John Passaniti

unread,
Feb 2, 2008, 3:55:55 PM2/2/08
to
Brad Eckert wrote:
> Between GPL and CDDL, which license would better serve the Open Source
> Forth community? GPL seems too strong and CDDL too weak.

I continue to fail to see why this is a Forth question. Are you
suggesting there is some attribute about Forth or Forth source code that
drives some need for a Forth-specific open source license?

When you look at other language communities, you see that there is no
monolithic thinking in their communities regarding licensing. You see
some code released under liberal licenses, other code under conservative
licenses, code released as public domain, and code with indeterminate
licenses because the authors don't care or don't understand the issue.

In other words, the question of licensing is pushed to the author, not
treated as something the entire community must agree to. It then
becomes the responsibility of the user of that code to follow the
license's restrictions.

Brad Eckert

unread,
Feb 4, 2008, 9:57:08 AM2/4/08
to
On Feb 2, 1:55 pm, John Passaniti <put-my-first-name-

h...@JapanIsShinto.com> wrote:
> Brad Eckert wrote:
> > Between GPL and CDDL, which license would better serve the Open Source
> > Forth community? GPL seems too strong and CDDL too weak.
>
> I continue to fail to see why this is a Forth question.  Are you
> suggesting there is some attribute about Forth or Forth source code that
> drives some need for a Forth-specific open source license?

It's a copyleft question, although Forth has some interesting side
effects. For example, if the source code uses a non-standard dialect
(it runs on a custom Forth) then GPL seems to require that the custom
Forth needs to be distributed too.

> In other words, the question of licensing is pushed to the author, not
> treated as something the entire community must agree to.  It then
> becomes the responsibility of the user of that code to follow the
> license's restrictions.

I wouldn't expect consensus in a forum like this. But I can stick my
finger up to see which way the wind blows. It seems to me GPL isn't a
big impediment to commercial developers. They can buy a non-GPL
license if necessary, and if contribution to the GPL code pool results
in reduced pricing of the non-GPL license then that's some nice
leverage.

Brad

Anton Ertl

unread,
Feb 4, 2008, 10:15:06 AM2/4/08
to
Brad Eckert <nospaa...@tinyboot.com> writes:
>For example, if the source code uses a non-standard dialect
>(it runs on a custom Forth) then GPL seems to require that the custom
>Forth needs to be distributed too.

I have recently discussed this with Brett Smith from the FSF, and they
view the interpreted program as constituting input for the
interpreter, not as deriving from the interpreter. They also think
that most programmers see it that way, and that judges will take that
into account when making a ruling.

See also
<http://www.fsf.org/licensing/licenses/gpl-faq.html#InterpreterIncompat>
and
<http://www.fsf.org/licensing/licenses/gpl-faq.html#IfInterpreterIsGPL>.

Stephen Pelc

unread,
Feb 4, 2008, 11:07:20 AM2/4/08
to
On Mon, 4 Feb 2008 06:57:08 -0800 (PST), Brad Eckert
<nospaa...@tinyboot.com> wrote:

>I wouldn't expect consensus in a forum like this. But I can stick my
>finger up to see which way the wind blows. It seems to me GPL isn't a
>big impediment to commercial developers. They can buy a non-GPL
>license if necessary, and if contribution to the GPL code pool results
>in reduced pricing of the non-GPL license then that's some nice
>leverage.

I am told, by people who claim to know, that the GPL problem is
real. For example, if you interface directly to the MySQL shared
library, the accessor should be under GPL. If you interface
through ODBC, the accessor does not need to be under GPL. This
situation just seems silly. I would much prefer clarity.

Proprietary toolmakers need to absolutely certain that they
are not passing restrictions or constraints to their customers.

Brad Eckert

unread,
Feb 4, 2008, 12:22:34 PM2/4/08
to
On Feb 4, 8:15 am, an...@mips.complang.tuwien.ac.at (Anton Ertl)
wrote:

> I have recently discussed this with Brett Smith from the FSF, and they
> view the interpreted program as constituting input for the
> interpreter, not as deriving from the interpreter.  They also think
> that most programmers see it that way, and that judges will take that
> into account when making a ruling.
>
Isn't the point of "Free Software" to ensure that the end user will
have the ability to change the code and recompile to fit new needs? If
I have a GPL calculator program and find that I never use the
factorial key but often need to do a cube root, I should be able to
change the function of that button to cube root instead of factorial
and then recompile the program.

If that program is Forth running on a custom interpreter (a virtual
machine not that conceptually different from the kind implemented with
gates and transistors) then a tool needs to be available to convert
the source code into a form executable by the VM. If the Forth does
this conversion when you INCLUDE the program, that's fine. But if the
code exists as part of a precompiled binary, where does the user get
the tools to recompile the calculator program?

Forth presents a sticky situation in that you can roll your own Forth
and not have to distribute it along with the source of your GPL-
compliant application. So yeah, the source code is out there for the
user to modify at her whim. She just has to write her own tools
(working from little or no documentation) to recompile it.

Brad

Anton Ertl

unread,
Feb 4, 2008, 1:40:48 PM2/4/08
to
Brad Eckert <nospaa...@tinyboot.com> writes:
>On Feb 4, 8:15=A0am, an...@mips.complang.tuwien.ac.at (Anton Ertl)

>wrote:
>> I have recently discussed this with Brett Smith from the FSF, and they
>> view the interpreted program as constituting input for the
>> interpreter, not as deriving from the interpreter. =A0They also think

>> that most programmers see it that way, and that judges will take that
>> into account when making a ruling.
...

>If that program is Forth running on a custom interpreter (a virtual
>machine not that conceptually different from the kind implemented with
>gates and transistors) then a tool needs to be available to convert
>the source code into a form executable by the VM. If the Forth does
>this conversion when you INCLUDE the program, that's fine.

That's the case with proper Forth systems. But the problem you point
out still exists: If only a proprietary Forth system can run the
program, every user needs to proprietary system to run it.

> But if the
>code exists as part of a precompiled binary, where does the user get
>the tools to recompile the calculator program?

From where the original user got it. If the tools are proprietary
software, then the original developers have trapped the free software
they developed.

That problem exists in other settings, too; actually it is more acute
in other settings; e.g., it used to be a problem with Java, and
therefore this problem has been called "The Java Trap" by the FSF.

The solution is to just avoid walking into the trap: only rely on free
tools (whether GPLed or licensed with a different free license).
However, nothing in the GPL requires that a program be processed only
with GPLed tools.

>Forth presents a sticky situation in that you can roll your own Forth
>and not have to distribute it along with the source of your GPL-
>compliant application. So yeah, the source code is out there for the
>user to modify at her whim. She just has to write her own tools
>(working from little or no documentation) to recompile it.

What would be the point of that? If you want to keep your code
proprietary, what would be the point of freeing a part of it, and if
you want to free your code, what would be the point of keeping a part
of it proprietary. Anyway, you can do so, but the value of the free
code will be close to zero, and most people will treat it as just
another proprietary program.

BTW, even assuming you were in violation of the GPL by doing this sort
of thing (because the two programs are really two parts of one
program), you have the copyright on both parts, so you are the only
one who has legal standing to sue you for doing this.

John Doty

unread,
Feb 4, 2008, 2:09:57 PM2/4/08
to
Brad Eckert wrote:

> Forth presents a sticky situation in that you can roll your own Forth
> and not have to distribute it along with the source of your GPL-
> compliant application. So yeah, the source code is out there for the
> user to modify at her whim. She just has to write her own tools
> (working from little or no documentation) to recompile it.

It's not peculiar to Forth. Today I'm writing some gloss around some
Mathematica code that I intend to distribute freely (although I'll
probably distribute it under CC-SA rather than GPL, because the gloss is
more important than the code, and that seems a more appropriate
license). Anyone who wants to actually run the code as is will need to
get Mathematica. Or rewrite it: the code is trivial. It's the idea
behind the code that's new.

Remember, the GPL governs distribution, not use. The end user may use
GPL software freely with any other software. The distributor, however,
is constrained by the GPL. Software distributed with GPL software must
be licensed under the GPL or a compatible license. But separate
distribution is not prohibited, and the GPL cannot endow the user with
any special rights to IP not covered by it.

--
John Doty, Noqsi Aerospace, Ltd.
http://www.noqsi.com/
--
History teaches that logical consistency is neither sufficient nor
necessary to establish practical, real world truth. Those who attempt to
use logic for that purpose are abusing it.

Brad Eckert

unread,
Feb 4, 2008, 4:13:06 PM2/4/08
to
On Feb 4, 11:40 am, an...@mips.complang.tuwien.ac.at (Anton Ertl)
wrote:
>

> What would be the point of that?

To be sneaky. Not me personally, but in theory it could be done.

> If you want to keep your code
> proprietary, what would be the point of freeing a part of it, and if
> you want to free your code, what would be the point of keeping a part
> of it proprietary.

Perhaps the same point as tivoization. You observe the letter of the
GPL law (linking to that nice GPL code requires publication) but end
users find the openly published code useless.

Brad

Jean-François Michaud

unread,
Feb 4, 2008, 4:55:14 PM2/4/08
to
On Feb 4, 8:07 am, stephen...@mpeforth.com (Stephen Pelc) wrote:
> On Mon, 4 Feb 2008 06:57:08 -0800 (PST), Brad Eckert
>
> <nospaambr...@tinyboot.com> wrote:
> >I wouldn't expect consensus in a forum like this. But I can stick my
> >finger up to see which way the wind blows. It seems to me GPL isn't a
> >big impediment to commercial developers. They can buy a non-GPL
> >license if necessary, and if contribution to the GPL code pool results
> >in reduced pricing of the non-GPL license then that's some nice
> >leverage.
>
> I am told, by people who claim to know, that the GPL problem is
> real. For example, if you interface directly to the MySQL shared
> library, the accessor should be under GPL. If you interface
> through ODBC, the accessor does not need to be under GPL. This
> situation just seems silly. I would much prefer clarity.

Only if you distribute. If the software is used internally and is
never distributed, you can do whatever the hell you want.

> Proprietary toolmakers need to absolutely certain that they
> are not passing restrictions or constraints to their customers.

They always are, but in different forms. For proprietary software, the
restriction is usually monetary and a non distribution agreement
whereas for open sourced code under GPL, the restriction is ensuring
source availability if distribution occurs (maybe a monetary
compensation as well, but not necessarilly).

The idea of the GPL is to make as much code available to the general
public so that end users end up being able to use almost anything at
almost no cost. It creates a pool of useful tools that can be combined
'ad nauseam' by whoever wants to put the effort into it. We can only
go up from there.

You can either feel comfortable floating about in the new thought
scheme or not. If in general, you don't feel that having access to
open source software is useful or that it can be value added to your
own development, then you can simply develop a proprietary solution in-
house and go about using the commonly accepted and known model using
known restrictions (usually monetary and non-distribution agreement).

If you feel that having access to open source code is value added and
will speed up your development considerably (it usually does) and are
ready to ensure availability on distribution, you simply have to
understand the restrictions which are fundamentally different from the
more direct and less open model.

Heck, if you don't distribute, you can build a company that
exclusively sits on open source software to run and not have to make
anything available since you aren't distributing a darn thing. If the
effort is worthwhile and useful, you can even decide to sell it as an
out of the box, ready to use business solution; at which point, the
higher level solution simply enters into the pool of publicly
available solutions that others can use and expand upon.

If everybody does the same thing, then everybody benefits from
everybody's developments and the cost somewhat equalizes across the
board. Everybody benefits while effectively making the cost of
acquiring source **code** solutions drop considerably (converging
towards zero). In-House Integration cost is not nil on the other hand,
quite the contrary :). Companies can focus their energy on integration
instead of in-house code development and maintenance. Businesses can
focus on selling already integrated solutions to other businesses and
check this out, the integrated solutions do not necessarily involve
source code modifications which means that if a solution simply
configures multiple open source components so they work in concert,
your solution, from my latest understanding, does not fall under the
GPL! This of course means that it can be sold, full price, under the
more commonly known and direct, less open, model.

Regards
Jean-Francois Michaud

Albert van der Horst

unread,
Feb 4, 2008, 5:30:12 PM2/4/08
to
In article <s4idnYpZcdqa_jra...@wispertel.com>,
John Doty <j...@whispertel.LoseTheH.net> wrote:
<SNIP>

>
>It's not peculiar to Forth. Today I'm writing some gloss around some
>Mathematica code that I intend to distribute freely (although I'll
>probably distribute it under CC-SA rather than GPL, because the gloss is
>more important than the code, and that seems a more appropriate
>license). Anyone who wants to actually run the code as is will need to
>get Mathematica. Or rewrite it: the code is trivial. It's the idea
>behind the code that's new.

I think this is an important point. There is merit to distributing
a program under GPL (or whatever open license) because the source
code has value even if you have no system that can compile and
run it as is. It can be adapted, implemented in ADA etc.
This is in contrast with e.g. a proprietary Nintendo program where you
have only the binary, maybe obfuscated and encrypted.
If the Nintendo boxes have died out, the only thing left are zero's
and ones.

<SNIP>

>
>--
>John Doty, Noqsi Aerospace, Ltd.

N.B. You have almost convinced me to drop Forth in favor of
Python :-)

Groetjes Albert

John Doty

unread,
Feb 4, 2008, 6:44:17 PM2/4/08
to
Albert van der Horst wrote:
> ...

> N.B. You have almost convinced me to drop Forth in favor of
> Python :-)

Stay at "almost". I really think they have different application
domains. But I think we Forthers can learn from Python's strengths.

--
John Doty, Noqsi Aerospace, Ltd.

Stephen J. Bevan

unread,
Feb 5, 2008, 1:49:11 AM2/5/08
to
steph...@mpeforth.com (Stephen Pelc) writes:
> I am told, by people who claim to know, that the GPL problem is
> real. For example, if you interface directly to the MySQL shared
> library, the accessor should be under GPL. If you interface
> through ODBC, the accessor does not need to be under GPL. This
> situation just seems silly. I would much prefer clarity.

Silly or not it is clear if you take into account the Unix heritage:
if two piences of code are in the same address space (whether static
or dynamic linking) and one is under the GPL then it all should be
under the GPL, if two pieces of code are in separate address spaces
then each address space can have different licenses and each can call
the other (using pipes, sockets, ... etc.) without restriction.

One can argue that the above is an arbitrary, or sven silly,
distinction that doesn't even exist in some environments (e.g. writing
a traditional monolithic Unix kernel!, but within the environment of
where it was created I think it is clear.


> Proprietary toolmakers need to absolutely certain that they
> are not passing restrictions or constraints to their customers.

If one wants absolute certaintity: ask the author(s) for a copy under
another license ($$$ sometimes helps) or don't use the GPL code.

Stephen Pelc

unread,
Feb 5, 2008, 5:33:01 AM2/5/08
to
On Mon, 4 Feb 2008 13:55:14 -0800 (PST),
"=?ISO-8859-1?Q?Jean-Fran=E7ois_Michaud?=" <com...@comcast.net>
wrote:

>The idea of the GPL is to make as much code available to the general
>public so that end users end up being able to use almost anything at
>almost no cost.

Stallman's consultancy rates are in dollars per minute.

It's a different business model, but it's still a business model,
and that business model needs clarity. From observation, a number
of businesses are moving away from GPL products because of the
lack of clarity in the GPL licenses.

Stephen Pelc

unread,
Feb 5, 2008, 6:34:32 AM2/5/08
to
On Tue, 05 Feb 2008 06:49:11 GMT, ste...@dino.dnsalias.com (Stephen
J. Bevan) wrote:

>Silly or not it is clear if you take into account the Unix heritage:
>if two piences of code are in the same address space (whether static
>or dynamic linking) and one is under the GPL then it all should be
>under the GPL, if two pieces of code are in separate address spaces
>then each address space can have different licenses and each can call
>the other (using pipes, sockets, ... etc.) without restriction.

Now that provides the clarity I needed - thank you. However, our
"ship's lawyer" says that the issue is about dependency on a
specific piece of code (MySQL), but the isolation provided by
ODBC removes the dependency.

If you have a dependency on a GPL item, then the accessor is a
derived work, and must follow the GPL.

In practice, both reasonings rule out MySQL and enable use of
SQLite and Postgres. It's a shame then that in these instances
the GPL is a competitive disadvantage.

Mark W. Humphries

unread,
Feb 5, 2008, 6:44:44 AM2/5/08
to
On Feb 5, 6:33 pm, stephen...@mpeforth.com (Stephen Pelc) wrote:
> On Mon, 4 Feb 2008 13:55:14 -0800 (PST),
> "=?ISO-8859-1?Q?Jean-Fran=E7ois_Michaud?=" <come...@comcast.net>

> wrote:
>
> >The idea of the GPL is to make as much code available to the general
> >public so that end users end up being able to use almost anything at
> >almost no cost.
>
> Stallman's consultancy rates are in dollars per minute.
>
> It's a different business model, but it's still a business model,
> and that business model needs clarity. From observation, a number
> of businesses are moving away from GPL products because of the
> lack of clarity in the GPL licenses.

We're seriously considering moving to FreeBSD because of this.

> Stephen
>
> --
> Stephen Pelc, stephen...@mpeforth.com

Bruce McFarling

unread,
Feb 5, 2008, 7:27:23 AM2/5/08
to
On Feb 5, 5:33 am, stephen...@mpeforth.com (Stephen Pelc) wrote:
> On Mon, 4 Feb 2008 13:55:14 -0800 (PST),
> "=?ISO-8859-1?Q?Jean-Fran=E7ois_Michaud?=" <come...@comcast.net>
> wrote:

> >The idea of the GPL is to make as much code available to the general
> >public so that end users end up being able to use almost anything at
> >almost no cost.

> Stallman's consultancy rates are in dollars per minute.

> It's a different business model, but it's still a business model,
> and that business model needs clarity. From observation, a number
> of businesses are moving away from GPL products because of the
> lack of clarity in the GPL licenses.

Yes, this gets to my original point. I recall it being argued in the
possibly distant past whether in some not-unusual configuration of
Forth system where there is no functional difference between the GPL
and the LGPL ... that is, whether the LGPL does not really allow the
Forth system to be used as a permissive Open Source library, but is
just as restrictive-to-open-source as the GPL. And both sides of the
argument had what seemed to be plausible cases.

Of course, if you are aiming at permissive Open Source, the most
straightforward solution may be to release under LGPL and BSD and let
whomever may wish to modify it and pass it on decide which one they
prefer.

Bruce McFarling

unread,
Feb 5, 2008, 7:34:32 AM2/5/08
to
On Feb 5, 6:34 am, stephen...@mpeforth.com (Stephen Pelc) wrote:
> In practice, both reasonings rule out MySQL and enable use of
> SQLite and Postgres. It's a shame then that in these instances
> the GPL is a competitive disadvantage.

For someone trying to propagate the GPL, if MySQL can provide a
competitive advantage to some GPL'd code that SQLite and Postgres
cannot provide, then that would be the point.

Of course, some code may be part of that process of trying to
propagate the GPL not because they have any such inclination, but
rather because the most aggressively open-source licensed library that
they link to is GPL'd.

Richard Owlett

unread,
Feb 5, 2008, 7:56:44 AM2/5/08
to
Stephen Pelc wrote:

> Jean-François Michaud wrote:
>
>
>>The idea of the GPL is to make as much code available to the general
>>public so that end users end up being able to use almost anything at
>>almost no cost.
>
>
> Stallman's consultancy rates are in dollars per minute.
>
> It's a different business model, but it's still a business model,
> and that business model needs clarity. From observation, a number
> of businesses are moving away from GPL products because of the
> lack of clarity in the GPL licenses.
>
> Stephen
>

Yeah BUT.
Does Mr. Stallman really support the right of a
craftsman/inventor/programmer/artist/author/... having the fruit of his
labor? I don't think so. He denies the concept of "Intellectual Property"

In early 2006 a vehement copyleft supporter wrote something that "got
under my skin" and referred me a page where Mr. Stallman expressed his
views - they got further under my skin.

So I emailed him with a subject line of "why NO SALE FSF/GNU/copyleft to
this right wing conservative entrepreneur". I didn't really expect a
reply but chose a title that would at least get read. He replied to my
message and a follow up.

He summarized by replying to my last question, " *How* does your system
reward/compensate me for Intellectual Property?"

Richard Stallman wrote:
>
> The term "intellectual property" embodies confusion and bias,
> so that use of that term is an obstacle to careful thought.
> See http://www.gnu.org/philosophy/not-ipr.xhtml for more explanation.

I just re-read that page. I think he uses biased wording at least as
much as he claims for his opponents. But it is worth reading as it
skillfully states his point of view.

>
> The term "compensate" also makes the assumption that something was
> unjustly taken from you. That doesn't relate to my views; I am not
> proposing to force you to write the software, or force you to release
> it.
>
> Your question, with the confusing terms removed is this: do I propose
> to reward you for writing a program that solves the problem?
>
> First I would ask, does that deserve a reward?
>
> To offer the public non-free software, software that citizens can use
> only at the cost of being under the developer's power, is evil -- it
> deserves a punishment. Under today's misguided system, it is likely
> to get a reward, but that is unjust and foolish; rewarding
> mistreatment of others encourages more mistreatment. I want to change
> this unjust system.
>
> To offer the public free software deserves a reward. I cannot promise
> you a reward if you do this, but there is a reasonable chance you may
> eventually get some sort of reward through it (such as employment, or
> professional reputation).
>
> You would also have the reward of knowing you contributed to making a
> better world. That reward is very important for me.
>
>
>

Brad Eckert

unread,
Feb 5, 2008, 10:08:09 AM2/5/08
to
On Feb 5, 5:56 am, Richard Owlett <rowl...@atlascomm.net> wrote:

>
> Richard Stallman wrote:
>  > To offer the public non-free software, software that citizens can use
>  > only at the cost of being under the developer's power, is evil -- it
>  > deserves a punishment.  Under today's misguided system, it is likely
>  > to get a reward, but that is unjust and foolish; rewarding
>  > mistreatment of others encourages more mistreatment.  I want to change
>  > this unjust system.

I'm starting to see why GPL is the way it is. But IMHO good ethics are
better promoted by positive reinforcement than by punishment. The
heavy-handed "Copyleft or else..." approach can be circumvented by
clever people, especially since Forth offers more flexibility for
doing such things. But then there's the question, "How long can we get
away with that?"

Embedded systems have a long history of secrecy. Only company insiders
know what's in Flash or masked ROM, and they like it that way. I would
like to encourage them to contribute to a common tool base without
forcing unreasonable demands on them the way GPL does. So far, CDDL
seems like the best option.

Brad

sp...@controlq.com

unread,
Feb 5, 2008, 10:42:36 AM2/5/08
to
On Tue, 5 Feb 2008, Stephen Pelc wrote:
> Stallman's consultancy rates are in dollars per minute.

I'm glad *I'm* not paying Stallman to release the Hurd then. 8-)
Ca-Ching!
---- Posted via Pronews.com - Premium Corporate Usenet News Provider ----
http://www.pronews.com offers corporate packages that have access to 100,000+ newsgroups

Jean-François Michaud

unread,
Feb 5, 2008, 11:33:37 AM2/5/08
to
On Feb 5, 2:33 am, stephen...@mpeforth.com (Stephen Pelc) wrote:
> On Mon, 4 Feb 2008 13:55:14 -0800 (PST),
> "=?ISO-8859-1?Q?Jean-Fran=E7ois_Michaud?=" <come...@comcast.net>

> wrote:
>
> >The idea of the GPL is to make as much code available to the general
> >public so that end users end up being able to use almost anything at
> >almost no cost.
>
> Stallman's consultancy rates are in dollars per minute.

There is still confusion between 'free' as in freedom and 'free' as in
free from fee. It is NOT a valid argument against the open source
model since it never actually claims to offer 'free' software in the
free from fee sense (it certainly is an interesting collateral effect
but it isn't the essence of open source). The open source model claims
software freedom. When we talk of free software, we don't talk about
it being free from fee. Period. People still have to make a living in
the current model. The context is important, otherwise, taken
literally without context, it is easy to confuse the two. Many are
still confused.

> It's a different business model, but it's still a business model,
> and that business model needs clarity. From observation, a number
> of businesses are moving away from GPL products because of the
> lack of clarity in the GPL licenses.

I'm not sure what isn't clear to you, spending a few hours on the net
to untangle everything is usually sufficient. Or at least, sufficient
enough to pull together a relatively good understanding.

Regards
Jean-Francois Michaud

Andrew Haley

unread,
Feb 5, 2008, 12:10:44 PM2/5/08
to
Richard Owlett <row...@atlascomm.net> wrote:
> Yeah BUT.
> Does Mr. Stallman really support the right of a
> craftsman/inventor/programmer/artist/author/... having the fruit of his
> labor? I don't think so.

Sure he does. He wants you to use free software, and to write free
software. He knows that you have a choice.

> He denies the concept of "Intellectual Property"

He actually says that trademark, copyright, and patent laws are very
different and have different goals, and it is confusing and misleading
to lump them together under the blanket term "Intellectual Property".
He's right about that.

Andrew.

Stephen Pelc

unread,
Feb 5, 2008, 12:43:52 PM2/5/08
to
On Tue, 5 Feb 2008 08:33:37 -0800 (PST),
"=?ISO-8859-1?Q?Jean-Fran=E7ois_Michaud?=" <com...@comcast.net>
wrote:

>There is still confusion between 'free' as in freedom and 'free' as in


>free from fee. It is NOT a valid argument against the open source
>model since it never actually claims to offer 'free' software in the

>free from fee sense ...

I'm not confused - I just quoted your previous post.

>> On Mon, 4 Feb 2008 13:55:14 -0800 (PST),
>> "=?ISO-8859-1?Q?Jean-Fran=E7ois_Michaud?=" <come...@comcast.net>
>> wrote:
>> >The idea of the GPL is to make as much code available to the general
>> >public so that end users end up being able to use almost anything at
>> >almost no cost.

------------^^^^

Aleksej Saushev

unread,
Feb 5, 2008, 1:26:57 PM2/5/08
to
"Jean-François Michaud" <com...@comcast.net> writes:

> On Feb 5, 2:33 am, stephen...@mpeforth.com (Stephen Pelc) wrote:
>> On Mon, 4 Feb 2008 13:55:14 -0800 (PST),
>> "=?ISO-8859-1?Q?Jean-Fran=E7ois_Michaud?=" <come...@comcast.net>
>> wrote:
>>
>> >The idea of the GPL is to make as much code available to the general
>> >public so that end users end up being able to use almost anything at
>> >almost no cost.
>>
>> Stallman's consultancy rates are in dollars per minute.
>
> There is still confusion between 'free' as in freedom and 'free' as in
> free from fee. It is NOT a valid argument against the open source
> model since it never actually claims to offer 'free' software in the
> free from fee sense (it certainly is an interesting collateral effect
> but it isn't the essence of open source). The open source model claims
> software freedom.

What "freedom" is it then?

If it were free, I could use the software _without_any_limitations_,
but Stallman imposes such conditions, that I don't even understand,
where the border lies, without prior consultation to very skillful
lawyer.

>> It's a different business model, but it's still a business model,
>> and that business model needs clarity. From observation, a number
>> of businesses are moving away from GPL products because of the
>> lack of clarity in the GPL licenses.
>
> I'm not sure what isn't clear to you, spending a few hours on the net
> to untangle everything is usually sufficient. Or at least, sufficient
> enough to pull together a relatively good understanding.

It's not clear to me, how can I distribute a bundle, containing
logically and mechanically separate GPL parts. The most logical
way is to list parts in supplementary documentation (because
only real geeks read it), providing URLs, where to get the source
(because end user doesn't want it in most of the case).
And it isn't clear, how the case of separate application with
command line interface is so much different from dynamic loaded
library with binary interface. Technically and logically those
are separate parts, hence linking to GPL library is allowed
without any restriction, when the library comes as separate file.
Why LGPL exists then?


--
CKOPO BECHA...
CKOPO CE3OH...

Brad Eckert

unread,
Feb 5, 2008, 2:42:36 PM2/5/08
to
On Feb 5, 9:33 am, "Jean-François Michaud" <come...@comcast.net>
wrote:

> When we talk of free software, we don't talk about
> it being free from fee. Period. People still have to make a living in
> the current model. The context is important, otherwise, taken
> literally without context, it is easy to confuse the two. Many are
> still confused.

It's another example of lack of clarity. If I charge a modest fee of a
million bucks for the source code, is it still considered "Free
Software"?

Stallman uses copyright law to his own ends, like making a round peg
fit in a square hole. It works but it's not pretty.

Brad

John Doty

unread,
Feb 5, 2008, 3:07:23 PM2/5/08
to
Stephen Pelc wrote:
> On Mon, 4 Feb 2008 13:55:14 -0800 (PST),
> "=?ISO-8859-1?Q?Jean-Fran=E7ois_Michaud?=" <com...@comcast.net>
> wrote:
>
>> The idea of the GPL is to make as much code available to the general
>> public so that end users end up being able to use almost anything at
>> almost no cost.
>
> Stallman's consultancy rates are in dollars per minute.
>
> It's a different business model, but it's still a business model,
> and that business model needs clarity.

Part of your conceptual difficulty is that you never use the word
"distribute" when you try to explain your confusion. GPL is
fundamentally a distribution license.

The business model is clear if you want to follow it. If you don't
really want to follow it, it becomes unclear where the line between
unethical and illegal behavior is. But that's a common difficulty with
law, not confined to the GPL.

> From observation, a number
> of businesses are moving away from GPL products because of the
> lack of clarity in the GPL licenses.

Yes, and they do this as their customers move toward GPL products. Hmmm...

Jean-François Michaud

unread,
Feb 5, 2008, 4:37:33 PM2/5/08
to
On Feb 5, 11:42 am, Brad Eckert <nospaambr...@tinyboot.com> wrote:
> On Feb 5, 9:33 am, "Jean-François Michaud" <come...@comcast.net>
> wrote:
>
> > When we talk of free software, we don't talk about
> > it being free from fee. Period. People still have to make a living in
> > the current model. The context is important, otherwise, taken
> > literally without context, it is easy to confuse the two. Many are
> > still confused.
>
> It's another example of lack of clarity. If I charge a modest fee of a
> million bucks for the source code, is it still considered "Free
> Software"?

Yes because the general public has the freedom to use it and to
reshape it however desired without having to answer to anybody; they
are free to use the software as they see fit for their own purposes
given they don't distribute after reworking the software. If
distribution occurs, under a license like GPL, then restrictions
apply. 'Open source' might resonate with you more, as opposed to
'closed source', usually associated with proprietary software. If
someone lays hands on source code for software that is proprietary
('closed source'; not free in the freedom sense), it usually spells
trouble.

In the free world, it's perfectly normal and accepted to acquire and
play with software source; given certain restrictions on derived work
in a distribution context.

> Stallman uses copyright law to his own ends, like making a round peg
> fit in a square hole. It works but it's not pretty.

But then again, so does everybody :). It's simply markedly different
from the common way of making copyright law work for one's ends.

Regards
Jean-Francois Michaud

Jean-François Michaud

unread,
Feb 5, 2008, 4:40:56 PM2/5/08
to
On Feb 5, 12:07 pm, John Doty <j...@whispertel.LoseTheH.net> wrote:
> Stephen Pelc wrote:
> > On Mon, 4 Feb 2008 13:55:14 -0800 (PST),
> > "=?ISO-8859-1?Q?Jean-Fran=E7ois_Michaud?=" <come...@comcast.net>

> > wrote:
>
> >> The idea of the GPL is to make as much code available to the general
> >> public so that end users end up being able to use almost anything at
> >> almost no cost.
>
> > Stallman's consultancy rates are in dollars per minute.
>
> > It's a different business model, but it's still a business model,
> > and that business model needs clarity.
>
> Part of your conceptual difficulty is that you never use the word
> "distribute" when you try to explain your confusion. GPL is
> fundamentally a distribution license.

One down.

> The business model is clear if you want to follow it. If you don't
> really want to follow it, it becomes unclear where the line between
> unethical and illegal behavior is. But that's a common difficulty with
> law, not confined to the GPL.

And another one.

> > From observation, a number
> > of businesses are moving away from GPL products because of the
> > lack of clarity in the GPL licenses.
>
> Yes, and they do this as their customers move toward GPL products. Hmmm...

Heheh, indeed.

Regards
Jean-Francois Michaud

John Doty

unread,
Feb 5, 2008, 4:42:36 PM2/5/08
to
Brad Eckert wrote:
> On Feb 5, 9:33 am, "Jean-François Michaud" <come...@comcast.net>
> wrote:
>> When we talk of free software, we don't talk about
>> it being free from fee. Period. People still have to make a living in
>> the current model. The context is important, otherwise, taken
>> literally without context, it is easy to confuse the two. Many are
>> still confused.
>
> It's another example of lack of clarity. If I charge a modest fee of a
> million bucks for the source code, is it still considered "Free
> Software"?

I think not. But you're still trying to figure out how to cheat. Stop
trying to do that and clarity will return. And you may indeed decide
that GPL isn't what you want. That's OK, but complaining that the GPL
doesn't let you do things that the copyright holders consider theft
won't get you anywhere. It's your privilege to choose license terms for
your IP. But if you want access to other's IP, you must abide by *their*
terms.

>
> Stallman uses copyright law to his own ends,

Who doesn't?

> like making a round peg
> fit in a square hole. It works but it's not pretty.

When was the last time you saw a "pretty" license agreement?

Stephen J. Bevan

unread,
Feb 5, 2008, 9:36:14 PM2/5/08
to
steph...@mpeforth.com (Stephen Pelc) writes:
> On Tue, 05 Feb 2008 06:49:11 GMT, ste...@dino.dnsalias.com (Stephen
> J. Bevan) wrote:
>
>>Silly or not it is clear if you take into account the Unix heritage:
>>if two piences of code are in the same address space (whether static
>>or dynamic linking) and one is under the GPL then it all should be
>>under the GPL, if two pieces of code are in separate address spaces
>>then each address space can have different licenses and each can call
>>the other (using pipes, sockets, ... etc.) without restriction.
>
> Now that provides the clarity I needed - thank you. However, our
> "ship's lawyer" says that the issue is about dependency on a
> specific piece of code (MySQL), but the isolation provided by
> ODBC removes the dependency.

The isolation is because the MySQL ODBC driver uses socket to connect
to the database thereby keeping a non-GPL application and the GPL
MySQL in separate addresse spaces. Given that the MySQL driver is
under the GPL then the separate address spaces would be moot if it was
not also for the fact that MySQL has an exception in its ODBC driver
license that allows it to be used in an non-GPL ODBC manager and thus
non-GPL program.


> If you have a dependency on a GPL item, then the accessor is a
> derived work, and must follow the GPL.

Without a defintion of dependency and accessor that sentence it is
unclear whether that sentence is true or not.

Stephen J. Bevan

unread,
Feb 5, 2008, 9:51:57 PM2/5/08
to
Aleksej Saushev <as...@inbox.ru> writes:
> What "freedom" is it then?
>
> If it were free, I could use the software _without_any_limitations_,

Then the only free code under your definition is that which has been
released into the public domain. Every other license contains one or
more restrictions/limitations, if it didn't it is equivalent to public
domain. For example the BSD license has some limitations. They
aren't particularly onerous, but they are limitations.

Stephen J. Bevan

unread,
Feb 5, 2008, 11:38:45 PM2/5/08
to
ste...@dino.dnsalias.com (Stephen J. Bevan) writes:

> steph...@mpeforth.com (Stephen Pelc) writes:
>> I am told, by people who claim to know, that the GPL problem is
>> real. For example, if you interface directly to the MySQL shared
>> library, the accessor should be under GPL. If you interface
>> through ODBC, the accessor does not need to be under GPL. This
>> situation just seems silly. I would much prefer clarity.
>
> Silly or not it is clear if you take into account the Unix heritage:
> if two piences of code are in the same address space (whether static
> or dynamic linking) and one is under the GPL then it all should be
> under the GPL, if two pieces of code are in separate address spaces
> then each address space can have different licenses and each can call
> the other (using pipes, sockets, ... etc.) without restriction.

In case it isn't obvious I'm not a lawyer and the above is my
interpretation only. Also some lawyers have argued that what I wrote
in the above is not true in some cases. Specifically if a non-GPL
program, let's call it N, is calling a GPL program, let's call it G,
over a pipe/socket and that there is no other program that you could
replace G with and have N still work then they argue N is derivative
of G and so falls under the GPL. I haven't personally spoken to such
lawyers and thus don't have an explanation as to why they would argue
this and at the same time argue that a user-space program running on
Linux that uses a system call that only exist under Linux (e.g. epoll)
is not derivative. Perhaps they'd just say the system call boundary
is special and that's an end to it. Not a satisfying answer from a
technological perspective, but then nor really is the distinction
between using a pipe/socket vs dynamic linking. Thus unless one is
willing to put all the code of a project under the GPL, the
simplest/safest thing is to not use any GPL code in the project.

Jean-François Michaud

unread,
Feb 6, 2008, 2:24:19 AM2/6/08
to
On Feb 5, 10:26 am, Aleksej Saushev <a...@inbox.ru> wrote:

> "Jean-François Michaud" <come...@comcast.net> writes:
> > On Feb 5, 2:33 am, stephen...@mpeforth.com (Stephen Pelc) wrote:
> >> On Mon, 4 Feb 2008 13:55:14 -0800 (PST),
> >> "=?ISO-8859-1?Q?Jean-Fran=E7ois_Michaud?=" <come...@comcast.net>
> >> wrote:
>
> >> >The idea of the GPL is to make as much code available to the general
> >> >public so that end users end up being able to use almost anything at
> >> >almost no cost.
>
> >> Stallman's consultancy rates are in dollars per minute.
>
> > There is still confusion between 'free' as in freedom and 'free' as in
> > free from fee. It is NOT a valid argument against the open source
> > model since it never actually claims to offer 'free' software in the
> > free from fee sense (it certainly is an interesting collateral effect
> > but it isn't the essence of open source). The open source model claims
> > software freedom.
>
> What "freedom" is it then?

It would probably be a good idea for you to read the full GPL license
if you want to understand. And I believe that we already covered the
broad lines of what it allows.

1- Use open source/GPL'd software for your own purpose and you can do
whatever you want. You can burn it to CD, set it on fire and tell the
whole world about it, you can sneeze on it, you can stare at it with
intensity, you can use the code as is or you can rewrite the whole
thing and decide to toss it or to use it, you can change a single byte
(adding a line feed character) or you can combine multiple such open
source pieces or a mix or proprietary and open source pieces to form a
useful solution and you don't have to answer to anybody (as far as the
open source code goes unde GPL goes anyways) AS LONG AS YOU DON'T
DISTRIBUTE YOUR SOLUTION.

2- If you distribute your solution, see the restrictions that apply to
derived work in the GPL license.

The clear result of this is that the flow is directed outwards, toward
the general public, as opposed to inwards toward the entities that
created the software (mostly corporate entities) by controlling the
flow of source code. Mr. Everybody can use such high quality solutions
for whatever purpose, usually for free (free from fee here), with the
freedom that comes through unlimited use, no conditions, as long as
the software is not modified for ****distribution****.

> If it were free, I could use the software _without_any_limitations_,

And you can, AS LONG AS YOU DON'T DISTRIBUTE DERIVED WORK.

> but Stallman imposes such conditions, that I don't even understand,
> where the border lies, without prior consultation to very skillful
> lawyer.
>
> >> It's a different business model, but it's still a business model,
> >> and that business model needs clarity. From observation, a number
> >> of businesses are moving away from GPL products because of the
> >> lack of clarity in the GPL licenses.
>
> > I'm not sure what isn't clear to you, spending a few hours on the net
> > to untangle everything is usually sufficient. Or at least, sufficient
> > enough to pull together a relatively good understanding.
>
> It's not clear to me, how can I distribute a bundle, containing
> logically and mechanically separate GPL parts. The most logical
> way is to list parts in supplementary documentation (because
> only real geeks read it), providing URLs, where to get the source
> (because end user doesn't want it in most of the case).
> And it isn't clear, how the case of separate application with
> command line interface is so much different from dynamic loaded
> library with binary interface. Technically and logically those
> are separate parts, hence linking to GPL library is allowed
> without any restriction, when the library comes as separate file.
> Why LGPL exists then?

The concepts/definitions that you are currently using to think about
this might not align well with the intended meaning of the wording
used in the license. You seem to want to use very sharp demarcations
whereas, in my experience, concepts are never that clean cut. From my
perception, conceptual boundaries are usually somewhat blurry. That's
why lawyers are so prolific ;-).

Regards
Jean-Francois Michaud

Brad Eckert

unread,
Feb 6, 2008, 9:13:36 AM2/6/08
to
> John Doty, Noqsi Aerospace, Ltd.http://www.noqsi.com/

Brad Eckert

unread,
Feb 6, 2008, 9:38:21 AM2/6/08
to
On Feb 5, 2:42 pm, John Doty <j...@whispertel.LoseTheH.net> wrote:
> I think not. But you're still trying to figure out how to cheat. Stop
> trying to do that and clarity will return. And you may indeed decide
> that GPL isn't what you want. That's OK, but complaining that the GPL
> doesn't let you do things that the copyright holders consider theft
> won't get you anywhere. It's your privilege to choose license terms for
> your IP. But if you want access to other's IP, you must abide by *their*
> terms.
>
I'm interested in building out a public Forth infrastructure, not
using somebody else's GPL code. The original question was: Which open
source license encourages the most code reuse?

Bernd suggested the dual-license GPL model (like Trolltech or MySQL).
If there are many contributors to the code pool besides myself (I can
dream, can't I?) then selling licenses may become complicated but
still manageable. If the code is primarily used in embedded systems,
most users will be forced out of the GPL model so the copyleft
incentive is gone. OTOH, I can bet that vendors will bend the rules
enough that they can publish their code, reap the benefits of GPL and
still keep important trade secrets in-house. At least then there is
copyleft incentive. Can you blame me for assuming that some businesses
play dirty?

Brad

none Byron Jeff

unread,
Feb 6, 2008, 11:25:09 AM2/6/08
to
In article <8763x3x...@inbox.ru>, Aleksej Saushev <as...@inbox.ru> wrote:
>"Jean-François Michaud" <com...@comcast.net> writes:
>
>> On Feb 5, 2:33 am, stephen...@mpeforth.com (Stephen Pelc) wrote:
>>> On Mon, 4 Feb 2008 13:55:14 -0800 (PST),
>>> "=?ISO-8859-1?Q?Jean-Fran=E7ois_Michaud?=" <come...@comcast.net>
>>> wrote:
>>>
>>> >The idea of the GPL is to make as much code available to the general
>>> >public so that end users end up being able to use almost anything at
>>> >almost no cost.
>>>
>>> Stallman's consultancy rates are in dollars per minute.
>>
>> There is still confusion between 'free' as in freedom and 'free' as in
>> free from fee. It is NOT a valid argument against the open source
>> model since it never actually claims to offer 'free' software in the
>> free from fee sense (it certainly is an interesting collateral effect
>> but it isn't the essence of open source). The open source model claims
>> software freedom.
>
>What "freedom" is it then?

Actually the more pertinent question is "freedom to who?"

>If it were free, I could use the software _without_any_limitations_,

Actually you can use the software without any limitations. The GPL and
the like impose rules on modification and distribution, not use.

>but Stallman imposes such conditions, that I don't even understand,
>where the border lies, without prior consultation to very skillful
>lawyer.

It's really not that complicated. The GPL (at least up to version 2) is
a reciprocity license. You receive a GPL licensed product with a certain
set of rights, and you can't abridge those rights to anyone that you
redistribute the product (or deriviatives). The short list:

1) You get the source.
2) You can modify the source.
3) You can redistribute the source. In fact you must redistribute the
source to anyone that you give/sell the product to. See #1 above.
4) You can charge for that redistribution.
5) You can use the software as you see fit.
6) You cannot bundle the software with incompatibly licensed software.

That's about it. All the complexity comes in when trying to mix those
rules with closed source software. Questions like:

"I received the source for a GPL program. Can I make changes and sell
the changed program without giving up the source for my changes?"

Violates rule #1 and #3.

All the nuances come in when trying to determine if something is a
derivative. Anything that isn't standalone, requires either
recompilation or relinking to change, is considered derivative. Most of
the derivative issue is based around rule #2.

Now the problem with the GPL is that it's pretty much an isolated
universe. If you want to do something that's incompatible with the list
above, then it just doesn't work. And in embedded systems work, #6 is
the biggest sticking point because all the software is in the same
space.

CDDL releases distributors from the #6 requirement. So it can play with
other licensed code.

>>> It's a different business model, but it's still a business model,
>>> and that business model needs clarity. From observation, a number
>>> of businesses are moving away from GPL products because of the
>>> lack of clarity in the GPL licenses.
>>
>> I'm not sure what isn't clear to you, spending a few hours on the net
>> to untangle everything is usually sufficient. Or at least, sufficient
>> enough to pull together a relatively good understanding.
>
>It's not clear to me, how can I distribute a bundle, containing
>logically and mechanically separate GPL parts. The most logical
>way is to list parts in supplementary documentation (because
>only real geeks read it), providing URLs, where to get the source
>(because end user doesn't want it in most of the case).

Actually the simplest way is to simply ship the source with the product
and be done with it. That releases you from any further obligation.

And just my opinion, it's not a good idea to deal with the end users
wants. The fact is down the line when they need something else from the
software, then they'll look for that source so that they can get someone
to make the mods for them (if they cannot do it themselves).

>And it isn't clear, how the case of separate application with
>command line interface is so much different from dynamic loaded
>library with binary interface.

One technical nuance alluded to above. Dynamic libraries are linked into
the application, even if it's ex post facto to the beginning of
execution. A command line program exists in a completely separate
process space, it can function standalone, and uses non process memory
for communication. So it's considered to be separate.

>Technically and logically those
>are separate parts, hence linking to GPL library is allowed
>without any restriction, when the library comes as separate file.

How is there no restriction? Even if you build a separate application
out of the GPL library and some CLI interface, the code for that
interface would have to be released. Now I agree that the application
that uses that CLI can be closed source. But nothing has impinged on the
integrity of the library (or the interfaces to it) itself. And that's
what GPL advocates are trying to prevent.

>Why LGPL exists then?

It's a reality check on the fact that the GPL is a closed universe. The
LGPL facilitates the building of free infrastructure that's protected
from being co-opted, while facilitating the use of that infrastructure
with code under other licenses. So it opens up libraries that are free
for usage with applications that are not.

But the LGPL devolves into the GPL when dealing with embedded systems
because there's only one address space for both infrastructure and
application. So you have to release source for both so that updates can
be made downstream.

The CDDL does not require that. So the end user does give up some rights
so that developers can retain the ability to more freely mix free and
non free code.

BAJ

none Byron Jeff

unread,
Feb 6, 2008, 11:29:25 AM2/6/08
to
In article <68010005-451d-4a27...@v46g2000hsv.googlegroups.com>,

Brad Eckert <nospaa...@tinyboot.com> wrote:
>On Feb 5, 9:33 am, "Jean-François Michaud" <come...@comcast.net>
>wrote:
>> When we talk of free software, we don't talk about
>> it being free from fee. Period. People still have to make a living in
>> the current model. The context is important, otherwise, taken
>> literally without context, it is easy to confuse the two. Many are
>> still confused.
>
>It's another example of lack of clarity. If I charge a modest fee of a
>million bucks for the source code, is it still considered "Free
>Software"?

Yup. Just as long as you give the source to the buyer.

Of course the buyer can then turn around and sell copies of the software
for whatever price they like. Again they are required to give source to
their buyers.

An open market will quickly drive the price down toward zero.

So I guess you'd better invest that million! ;-)

>
>Stallman uses copyright law to his own ends, like making a round peg
>fit in a square hole. It works but it's not pretty.

True. Read his book and he explains early on about how awful he felt
about being locked out of his own code. There is a driving force behind
his zealotry.

BAJ

none Byron Jeff

unread,
Feb 6, 2008, 11:36:23 AM2/6/08
to
In article <41946d1b-05eb-4aae...@i12g2000prf.googlegroups.com>,

Jean-François Michaud <com...@comcast.net> wrote:
>On Feb 5, 11:42 am, Brad Eckert <nospaambr...@tinyboot.com> wrote:
>> On Feb 5, 9:33 am, "Jean-François Michaud" <come...@comcast.net>
>> wrote:
>>
>> > When we talk of free software, we don't talk about
>> > it being free from fee. Period. People still have to make a living in
>> > the current model. The context is important, otherwise, taken
>> > literally without context, it is easy to confuse the two. Many are
>> > still confused.
>>
>> It's another example of lack of clarity. If I charge a modest fee of a
>> million bucks for the source code, is it still considered "Free
>> Software"?
>
>Yes because the general public has the freedom to use it and to
>reshape it however desired without having to answer to anybody;

Technically that's not true. The code is not obligated to go public as
the buyer is not obligated to release the code to anyone.

Each and every code holder has the right to redistribute (even for a
fee). But they are not obligated to do so.

So imagine a situation where you get a contract job to build a system
worth $4 million, but because of the existing GPL infrastructure you can
get it done for $1 million. But because the infrastructure is GPL, the
application needs to be GPLed to. You sell to the buyer for $1 million,
it's free software, but the buyer has no obligated to disclose that
source to anyone.

It's an obscure, yet valid situation.

>they
>are free to use the software as they see fit for their own purposes
>given they don't distribute after reworking the software. If
>distribution occurs, under a license like GPL, then restrictions
>apply.

That is correct. All the restructions are on redistribution.

BAJ

none Byron Jeff

unread,
Feb 6, 2008, 11:45:48 AM2/6/08
to
In article <VfCdneNp-9DGRTXa...@wispertel.com>,

John Doty <j...@whispertel.LoseTheH.net> wrote:
>Brad Eckert wrote:
>> On Feb 5, 9:33 am, "Jean-François Michaud" <come...@comcast.net>
>> wrote:
>>> When we talk of free software, we don't talk about
>>> it being free from fee. Period. People still have to make a living in
>>> the current model. The context is important, otherwise, taken
>>> literally without context, it is easy to confuse the two. Many are
>>> still confused.
>>
>> It's another example of lack of clarity. If I charge a modest fee of a
>> million bucks for the source code, is it still considered "Free
>> Software"?
>
>I think not.

Why not? Money isn't the objective of free software.

>But you're still trying to figure out how to cheat.

How? Free software can be sold. What's the problem?

>Stop trying to do that and clarity will return.

I think your spreading misconceptions. I find it muddies the waters.

Again the rules:

1) You can sell free software.
2) You must distribute source with that free software.
3) You cannot restrict anyone who receives free software from
redistributing it.

So selling it isn't cheating. Selling it without source and telling the
buyer that they cannot redistribute. Now that's cheating.

>And you may indeed decide
>that GPL isn't what you want. That's OK, but complaining that the GPL
>doesn't let you do things that the copyright holders consider theft
>won't get you anywhere. It's your privilege to choose license terms for
>your IP. But if you want access to other's IP, you must abide by *their*
>terms.

I think that's a bit harsh. Brad is just trying to understand how the
boundaries work.

And IIRC Brad started this thread asking what's the best license to use
(emphesis mine) HIS OWN IP THAT HE WANTED TO SHARE!

I know that many have problems with how the GPL, and other free software
licenses, operate. However, choosing the right one for the right
situation is important.

BAJ

none Byron Jeff

unread,
Feb 6, 2008, 11:49:23 AM2/6/08