[Q] GPL violation.
Paco Moya
It was about a device driver for a frame grabber which could be
dynamicaly linked with Linux or statically linked with it.
Some people, namely Linus himself, stated that it was fair use of
the GPL'd code (the kernel) because no line of code was apparently
used and the driver was independent enough.
I'm not a lawyer but I cannot accept this interpretation of code reuse.
A device driver is not in any way independent of the kernel with which
it interacts. I'm not talking about interface copyright or patents but
about logical dependance.
We're about to release a GPL'd library written in C++ using libdl for
easy expansion in runtime. The library implements a base class Filter
which provides support for making instances of derived classes loaded
in runtime. It sounds similar to the case of the proprietary driver
but we do not want proprietary Filters derived from our library.
Dynamic linking is just a special case but object oriented development
is a more general case where derived work don't actually use a single
line of code from libraries but use them. Why bother to write LGPL if
anybody may make works derived from GPL'd code without distributing
sources? If the legal interpretation of GPL allows such behaviours there
is no need for LGPL at all.
I don't know what does the GPL or Copyright law consider a derived work.
Could anyone explain wheather the GPL would allow proprietary classes
derived from a GPL'd class? Does dynamic linking make difference?
Thanx a lot,
Paco. fm...@die.upm.es
pa...@neptuno.die.upm.es
Ingemar Hulthage
Paco Moya (paco@I_should_put_my_domain_in_etc_NNTP_INEWS_DOMAIN) wrote:
...
> We're about to release a GPL'd library written in C++ using libdl for
> easy expansion in runtime. The library implements a base class Filter
> which provides support for making instances of derived classes loaded
> in runtime. It sounds similar to the case of the proprietary driver
> but we do not want proprietary Filters derived from our library.
Why not ? What are the goals that your are trying to accomplish by
allowing everyone to copy your code ? Are you sure that it would be
in conflict with your goals if somebody made a proprietary enhancement
to it ?
> Dynamic linking is just a special case but object oriented development
> is a more general case where derived work don't actually use a single
> line of code from libraries but use them. Why bother to write LGPL if
> anybody may make works derived from GPL'd code without distributing
> sources?
I'm not sure this answers your question, but it is legal to create
derivative work of GPL code and not distribute it at all, in any form.
This is, in my opinion, an undesirable consequence of the GPL. Since
it is impossible to retain exclusive rights to derivative works, some
such work might not be distributed at all, commercially or otherwise.
> I don't know what does the GPL or Copyright law consider a derived work.
I'm not an attorney, but I believe that a derived work is anything
that contains parts of another work in verbatim, but also paraphrased
or translated in some way.
Ingemar Hulthage
Ajit George
: paco@I_should_put_my_domain_in_etc_NNTP_INEWS_DOMAIN (Paco Moya) writes:
: >Some people, namely Linus himself, stated that it was fair use of
: >the GPL'd code (the kernel) because no line of code was apparently
: >used and the driver was independent enough.
: >I'm not a lawyer but I cannot accept this interpretation of code reuse.
: I'm not a lawyer either, but ...
: Doesn't matter. As the copyright owner Linus is free to grant extra
: permissions beyond what the GPL allows, so even if you're right, Linus's
: word goes.
That's assuming that the word in question was written into the license
distributed with the code. Linus can't make new wishes apply
retroactively.
Oh yeah, I'm not a lawyer either. Just picking nits.
Has anyone taken a case of GPL violation to court? We can speculate
all we like about technicalities of the GPL, but if some judge gets up
on his bench one day and says the whole thing is as worthless as the
paper it's printed on (or not, as the case may be), then the points
are all moot.
Ajit
Joe Buck
Check the above. -^
>Some weeks ago I posted an article on a probable violation of the GPL.
>It was about a device driver for a frame grabber which could be
>dynamicaly linked with Linux or statically linked with it.
>
>Some people, namely Linus himself, stated that it was fair use of
>the GPL'd code (the kernel) because no line of code was apparently
>used and the driver was independent enough.
>I'm not a lawyer but I cannot accept this interpretation of code reuse.
I'm not a lawyer either, but ...
Doesn't matter. As the copyright owner Linus is free to grant extra
permissions beyond what the GPL allows, so even if you're right, Linus's
word goes. (Some other person who's written a major piece of the kernel
would have standing to object, perhaps, but their case would be questionable).
>A device driver is not in any way independent of the kernel with which
>it interacts. I'm not talking about interface copyright or patents but
>about logical dependance.
How would you prove infringement when the alleged infringer is not
distributing even one line of GPL'ed code, and is using standard,
documented interfaces? This is the user-does-the-link question again,
but in the NeXT case (the Objective-C compiler) they were going to alter
gcc to fit with their piece, rather than just use a standard interface,
so I think that the law could make a distinction.
>We're about to release a GPL'd library written in C++ using libdl for
>easy expansion in runtime. The library implements a base class Filter
>which provides support for making instances of derived classes loaded
>in runtime. It sounds similar to the case of the proprietary driver
>but we do not want proprietary Filters derived from our library.
The GPL would protect you against anyone distributing a binary containing
both your code and their code, unless they put their code under the GPL.
If your header files do inline functions, they wouldn't be able to
compile their code separately and claim they aren't violating your
copyright, since it would be a derivative work of the header file.
>Could anyone explain wheather the GPL would allow proprietary classes
>derived from a GPL'd class? Does dynamic linking make difference?
They can write the source. If they compile the source, the object
file will be a derivative work of both their source and your header
file. They can't distribute the object without your permission.
They can't distribute a binary that links them both together.
They *could* distribute the *source* without your permission, it
seems, though the user-does-the-link flames will start when we
talk about this. (There was a big stink about use of the GNU MP
library with RIPEM on this subject).
Note that copyright doesn't treat derived classes in C++ any differently
from, say, a class that had your class as a member. Don't let the
term "derived class" confuse you, it has nothing to do with "derivative
work".
--
-- Joe Buck <jb...@synopsys.com> (not speaking for Synopsys, Inc)
Linus Torvalds
Paco Moya <paco@I_should_put_my_domain_in_etc_NNTP_INEWS_DOMAIN> wrote:
>Some weeks ago I posted an article on a probable violation of the GPL.
>It was about a device driver for a frame grabber which could be
>dynamicaly linked with Linux or statically linked with it.
>
>Some people, namely Linus himself, stated that it was fair use of
>the GPL'd code (the kernel) because no line of code was apparently
>used and the driver was independent enough.
>
>I'm not a lawyer but I cannot accept this interpretation of code reuse.
>it interacts. I'm not talking about interface copyright or patents but
>about logical dependance.
Note that there is no such thing as "dynamically link into the kernel"
in linux. Instead there are "loadable modules".
Now the above may strike some people as nit-picking, but there is one
rather important thing about loadable modules: they can _not_ link
themselves against any random kernel routine. And the routines they
_can_ link against are routines that I consider to be "logically
independent" of the kernel implementation.
Essentially, the kernel module interface is a "library" interface to the
kernel, and kernel modules are considered to be under the GNU _Library_
license. In fact, due to the way kernel modules work, you automatically
do it according to the LGPL so this isn't explicitly stated anywhere,
but that's the way you should think about this.
Another way to look at this - using the legal rather than the moral
viewpoint - is to just see module loading as "use" of the kernel, rather
than as linking against it. I prefer to explain the rationale behind it
using the _moral_ reason to do it, though.
The reason the kernel is exposed in such a LGPL'd way when using modules
is simply that there are a lot of kernel device drivers for unix
available, and they were not all written under linux. If somebody wants
to port his SVR4 driver to linux but doesn't want to GPL it, I feel that
he should have the right to do that, using modules. After all, the
driver wasn't actually derived from linux itself: it's a real driver in
its own right, so I don't feel that I have the moral right to force him
to switch copyrights.
Now, the above said, I _much_ _much_ prefer GPL'd drivers, even if they
are available only as modules. Especially if they were actually
originally written _for_ linux, I consider it a bit dodgy to not use the
GPL (they can potentially be considered derived works even if you don't
actually link them into the kernel per se). But I do not want to force
it on people that arguably are _not_ doing derived work (it would be
rather preposterous to call the Andrew FileSystem a "derived work" of
linux, for example, so I think it's perfectly ok to have a AFS module,
for example).
For several reasons a linux module also doesn't always make much sense
unless it comes with sources - if some commercial company thinks that
linux is important enough that they want to do a commercial module for
linux, they may also recognize that a binary module doesn't work for
most linux users who use experimental kernels, for example.
Final note: the linux interpretation is not a "normal" case. I wouldn't
use it as a guide-line to anything else, especially not in user mode.
Linus
Francisco Moya Fernandez
: Paco Moya (paco@I_should_put_my_domain_in_etc_NNTP_INEWS_DOMAIN) wrote:
: ...
: > We're about to release a GPL'd library written in C++ using libdl for
: > easy expansion in runtime. The library implements a base class Filter
: > which provides support for making instances of derived classes loaded
: > in runtime. It sounds similar to the case of the proprietary driver
: > but we do not want proprietary Filters derived from our library.
: Why not ? What are the goals that your are trying to accomplish by
: allowing everyone to copy your code ? Are you sure that it would be
: in conflict with your goals if somebody made a proprietary enhancement
: to it ?
: > Dynamic linking is just a special case but object oriented development
: > is a more general case where derived work don't actually use a single
: > line of code from libraries but use them. Why bother to write LGPL if
: > anybody may make works derived from GPL'd code without distributing
: > sources?
: I'm not sure this answers your question, but it is legal to create
: derivative work of GPL code and not distribute it at all, in any form.
: This is, in my opinion, an undesirable consequence of the GPL. Since
: it is impossible to retain exclusive rights to derivative works, some
: such work might not be distributed at all, commercially or otherwise.
: > I don't know what does the GPL or Copyright law consider a derived work.
: I'm not an attorney, but I believe that a derived work is anything
: that contains parts of another work in verbatim, but also paraphrased
: or translated in some way.
: Ingemar Hulthage
--
----------------------------- Escuela Tecnica Superior
Francisco Moya Fernandez de Ingenieros de Telecomunicacion.
pa...@neptuno.die.upm.es Universidad Politecnica de Madrid.
Francisco Moya Fernandez
Ingemar Hulthage (hult...@hollywood.cinenet.net) wrote:
> Paco Moya (pa...@neptuno.die.upm.es) wrote:
> >...
> > but we do not want proprietary Filters derived from our library.
> Why not ? What are the goals that your are trying to accomplish by
> allowing everyone to copy your code ?
I'll quote the GNU Manifesto:
I consider that the golden rule requires that if I like a program I
must share it with other people who like it. Software sellers want to
divide the users and conquer them, making each user agree not to share
with others. I refuse to break solidarity with other users in this
way. I cannot in good conscience sign a nondisclosure agreement or a
software license agreement.
At least another paragraph applies here:
All intellectual property rights are just licenses granted by
society because it was thought, rightly or wrongly, that society as a
whole would benefit by granting them. But in any particular
situation, we have to ask: are we really better off granting such
license? What kind of act are we licensing a person to do?
> Are you sure that it would be in conflict with your goals if somebody
> made a proprietary enhancement to it ?
Yes. I don't see any good reason to distribute it under a less
restrictive license (such as LGPL). If you don't care whether other
people make proprietary enhancements then you probably don't care
whether it is distributed under the GPL or not. The same benefits you
obtain from GPL'd code are retained by not allowing proprietary
enhancements.
> I'm not sure this answers your question, but it is legal to create
> derivative work of GPL code and not distribute it at all, in any form.
I can't avoid that. But I am able to avoid proprietary enhancements.
> This is, in my opinion, an undesirable consequence of the GPL. Since
> it is impossible to retain exclusive rights to derivative works, some
> such work might not be distributed at all, commercially or otherwise.
It is not undesirable for me in any way. When a program is distributed
under the GPL those consequences are just the intention of the author.
Francisco Moya Fernandez
> paco@I_should_put_my_domain_in_etc_NNTP_INEWS_DOMAIN (Paco Moya) writes:
> Check the above. -^
Sorry, I think it's already fixed :)
> Doesn't matter. As the copyright owner Linus is free to grant extra
> permissions beyond what the GPL allows, so even if you're right,
> Linus's word goes.
I don't object to Linus' words. But I think Linus' interpretation of
the GPL is exactly my interpretation of the LGPL.
> How would you prove infringement when the alleged infringer is not
> distributing even one line of GPL'ed code, and is using standard,
> documented interfaces?
I'm not a lawyer. But the need to link the library (or the kernel, or
any piece of software) to make a running binary should be enough.
> The GPL would protect you against anyone distributing a binary
> containing both your code and their code, unless they put their code
> under the GPL.
When the GPL'd code supports dynamic linking users may develop
derivatives without even recompiling original sources. Would be legal
distributing full sources for my library and their libraries in binary
form?
> If your header files do inline functions, they wouldn't be able to
> compile their code separately and claim they aren't violating your
> copyright, since it would be a derivative work of the header file.
It would be trivial to outline inlines at a small (or not so small but
affordable) performance loss.
> They can write the source. If they compile the source, the object
> file will be a derivative work of both their source and your header
> file. They can't distribute the object without your permission.
> They can't distribute a binary that links them both together.
It sounds to me like interfaces copyright. No hope :)
> Note that copyright doesn't treat derived classes in C++ any
> differently from, say, a class that had your class as a member.
> Don't let the term "derived class" confuse you, it has nothing to do
> with "derivative work".
If derived classes are not considered derivative works by the
copyright law any software system developed with good OO methods and
with reusability in mind might be used by anyone by adding a subsystem
for dynamic linking. Well, I don't actually know whether dynamic
linking cares or not but it is a way not to include original code in
the distributed binary.
In proprietary software industries there is no much difference in
making derivative works by including other people's code or by using
it because each copy of the derived software will bring money to the
original copyright holder anyway. For free software authors it means
proprietary software companies may distribute derivatives just
including the source for the original program.
Paco
Mike Stump
Paco Moya <paco@I_should_put_my_domain_in_etc_NNTP_INEWS_DOMAIN> wrote:
>A device driver is not in any way independent of the kernel with which
>it interacts.
What if the driver works with Linux, FreeBSD, NetBSD and Mach? What
if the driver only works with Linux today, but tomorrow it works with
another independent kernel?
>we do not want proprietary Filters derived from our library.
Put it in your licence, if you want. No one forces you to use the
GPL as is. The text of the GPL is public domain, if I recall.
>Dynamic linking is just a special case but object oriented development
>is a more general case where derived work don't actually use a single
>line of code from libraries but use them. Why bother to write LGPL if
>anybody may make works derived from GPL'd code without distributing
>sources? If the legal interpretation of GPL allows such behaviours there
>is no need for LGPL at all.
>
>I don't know what does the GPL or Copyright law consider a derived work.
What GPL or rms, or the FSF thinks is a derived work, is pretty
irrelevant. What matters is what a court has said, or what the PTO
has said. Rms has stated things in the past (RIPEM) that I do not
agree with in this respect.
>Could anyone explain wheather the GPL would allow proprietary classes
>derived from a GPL'd class? Does dynamic linking make difference?
I like the statement:
If to use something, the use of that something requires an interface,
merely using the required interface does not by itself create a
derived work.
There is some support for this in law. There is also bound to be some
non-support for this in law. It may have to be litigated to be
resolved for any particular instance of it. Sorry, welcome to the US
legal system.
Michael I. Bushnell, p/BSG
I will agree, in the general case, he cannot. For example, he cannot
now say, all people must now pay me $100 to have it, or they must
remove it. But, when considering granting extra rights to people, yes
he can. He can also put totally new terms on it, like say, you must
pay $100 for each copy you have, and enforce it, but that would not
apply to old copies, only new copies.
Let me point out, lest anyone get scared, that in the case of the FSF
this kind of thing would be impossible.
Nearly all GNU software was contributed by its authors to the FSF.
(That's why the copyright is in the name of the FSF.) They did so by
signing a contract with the FSF, which grants the FSF the copyright,
but prohibits the FSF from doing distribution under any but free
software terms.
That means that the FSF is contractually bound, to *jillions* of
people, in a way that prevents it from hoarding software in the
future.
Even if you trust the FSF board, this is still a good thing, because
if the FSF loses a lawsuit someday and its assets are seized by Big
Bad Nasty Software, Inc., then BBNS can't distribute it under non-free
software terms, so it still remains free.
Michael
Steve Dunham
: Joe Buck (jb...@synopsys.com) writes:
: > paco@I_should_put_my_domain_in_etc_NNTP_INEWS_DOMAIN (Paco Moya) writes:
: > Check the above. -^
: Sorry, I think it's already fixed :)
: > Doesn't matter. As the copyright owner Linus is free to grant extra
: > permissions beyond what the GPL allows, so even if you're right,
: > Linus's word goes.
: I don't object to Linus' words. But I think Linus' interpretation of
: the GPL is exactly my interpretation of the LGPL.
So what's the difference between loading a kernel module and loading
an executable file?
Steve
dun...@gdl.msu.edu
Mike Stump
Ajit George <ga...@cs.utexas.edu> wrote:
>Joe Buck (jb...@synopsys.com) wrote:
>: Doesn't matter. As the copyright owner Linus is free to grant extra
>: permissions beyond what the GPL allows, so even if you're right, Linus's
>: word goes.
>
>retroactively.
As copyright holder, yes he can (in this case). You fail to
understand basic copyright law.
Terry Lambert
] If derived classes are not considered derivative works by the
] copyright law any software system developed with good OO methods and
] with reusability in mind might be used by anyone by adding a subsystem
] for dynamic linking. Well, I don't actually know whether dynamic
] linking cares or not but it is a way not to include original code in
] the distributed binary.
There is, in fact, a significant difference between loading a
kernel module and dynamic linking against a shared library.
As the original author of the BSD LKM mechanism (and as the
autor of a BSD shared library mechanism in June of 93, which
was unfortunately rendered undistributable), I am *very* familiar
with the ideas involved in both.
The difference is that in a loaded kernel module, the module
symbols are all imported at load time. In the shared library
case, typically the data symbols are imported at link time,
and the text is imported at load time.
That is, the static data referenced by the library itself and
the agregate initialized data for the library are seperate.
For kernel modules, this means that proprietary driver modules
can be loaded into a GPL kernel, but that a GPL kernel can not
be distributed with them statically linked in unless you can
also distribute the driver source (this is a pain in the butt
if it happens to be your disk driver).
Similarly, it means that GPL'ed drivers can be loaded into a
non-GPL'ed kernel, but that the resulting kernel can not be
distributed if you don't also distribute the sources AND the
license is non-conflicting with the GPL (specifically, no
additional caluses, like the BSD "credit clause"). This is
also a pain in the butt if it happens to be your disk driver.
Actually, if the kernel itself was LGPL'ed, then the LGPL relink
clause would apply, and you would be fine with the commercial
driver, as long as you distributed it as a linkable .o file. I
have asked about this several times with regard to Linux, but
the overhead of changing distribution terms now that there are
so many contributors means that it's unlikely the Linux could
be shipped with commercial disk drivers without an impossible
amount of effort.
If the drivers were LGPL'ed, then the non-GPL'ed kernel could
be linked with the LGPL'ed drivers, as long as a link kit was
provided (and source for the LGPL'ed drivers, of course). This
would have the effect of making use of LGPL'ed drivers possible
under SCO, UnixWare, FreeBSD, and other OS's that have
conflicting clauses or choose a binary distribution format, and
since the drivers themselves are under LGPL, the improvement and
source provision clauses protect their "freedom" (in the GNU
sense of the word). I've asked about this several times as well,
also with no success.
Now shared libraries (dynamic linking):
The LGPL requires that the user be able to replace the library
with an updated version (this is the "relink clause" of the LGPL).
It turns out that dynamic linking, as implemented by most shared
library technologies, doesn't meet the terms of the LGPL. In
specific: it is not possible to replace a shared library with
another in certain circumstances.
Consider the library function "fooinit":
char namebuf[ 8];
void
fooinit( char *name)
{
if( namebuf[ 0])
return;
strncpy( namebuf, name, sizeof( namebuf));
namebuf[ sizeof( namebuf) - 1 ] = 0; /* maybe truncate*/
}
In the next rev of the library, this has been changed to
char namebuf[ 32];
[ ... ]
Now, using the new library with the old data (*exactly* what
happens with shared libraries, since the data is stored with
the post-link binary instead of in the library) could result
in a segmentation violation, and if it doesn't, it *will* result
in semi-random data corruption.
A program using a shared library that implements sharing this
way (the way Linux and FreeBSD and NetBSD and SunOS and SVR4
and Solaris and ... currently do) can not be said to comply with
the terms of the relink clause of the LGPL.
So if we go back to our kernel, we see that the GPL'ed kernel
fails this case with a *dynamically linked* module, but not
with a *loaded* module.
The same failure occurs in the BSD case using an LGPL'ed driver:
if it is *dynamically linked*, then like the application linked
with the LGPL'ed shared library, it can't guarantee that future
versions of the library will be usable (without a full link kit
even in the LGPL'ed case).
So Linus' "nitpick" (see his posting in this thread) is
*exactly* right, and he is *exactly* correct about where the
distinction needs to be drawn in the "GPL'ed kernel with
commercial driver" case.
Terry Lambert
te...@cs.weber.edu
---
Any opinions in this posting are my own and not those of my present
or previous employers.
Joe Buck
Ajit George <ga...@cs.utexas.edu> wrote:
>>That's assuming that the word in question was written into the license
>>distributed with the code. Linus can't make new wishes apply
>>retroactively.
m...@kithrup.com (Mike Stump) writes:
>As copyright holder, yes he can (in this case). You fail to
>understand basic copyright law.
Or rather, yes he could if he were the sole copyright owner. At present,
though, only about half the code in the Linux kernel was written by Linus
Torvalds; most of the device drivers as well as some pieces of the kernel
itself were written by others, and they did not assign copyright to Linus
(as the FSF asks its contributors to do), so the whole group of people
would have to agree before conditions on the code could change.
For a list, see the CREDITS file in the kernel distribution.
H.J. Lu
In this case, I would call it a bug. To me it is the same as you
change the API/ABI. When you do that, the old programs may fail
no matter if static or dynamic link, nor the source code is provided
or not. You have to change something to get the old programs working
again with the new library.
H.J.
Craig Milo Rogers
legal advice.]
In article <4aprrr$7...@sanson.dit.upm.es> paco@I_should_put_my_domain_in_etc_NNTP_INEWS_DOMAIN (Paco Moya) writes:
>I'm not a lawyer but I cannot accept this interpretation of code reuse.
>A device driver is not in any way independent of the kernel with which
>it interacts. I'm not talking about interface copyright or patents but
>about logical dependance.
Your furtune cookie reads: Learn to be more accepting about
those things you cannot change. :-)
Copyright infringement and "logical dependance", by which I
presume you mean runtime execution dependence, are two different
things. In analogy: copyright violation applies to compiletime macro
expansion, not runtime subroutine calls. In other words, if one
program looks like a compiletime macro translation of another, than a
Copyright infringement might have occured. If one program happens to
make runtime calls to a second program without a
macro-translation-like relationship, than there is no Copyright
infringement. In general, a "use" which isn't "source reuse"
(allowing for the ill-defined term "translation") isn't a Copyright
infringement (at present).
For the specific case of a kernel device driver, it *is*
possible to write device drivers with "OS-dependent" and
"OS-independent" parts, although it may be economically justifiable
only when you can hide critical information (eg, register definitions
or bootstrap code) inside the "OS-independent" part. If you make the
first OS-dependent part for a non-GPLed OS, then create the
OS-dependent part for your target GPLed OS, your OS-independent part
is clearly independent of the GPLed OS, per the usual interpretation
of Copywrite law. QED.
It becomes an exercise in nitpicking as to whether you can
non-infringingly distribute, say, a CDROM with: 1) a binary image of
the GPLed OS kernel with your driver linked in, 2) sources to the
GPLed OS, 3) sources to your OS-dependent driver code for this GPLed
OS, and 4) no sources to your OS-independent driver code. It would
probably require a court precedent to establish this one way or
another.
It certainly seems reasonable that you could non-infringingly
distribute a CDROM with: 1) a binary image of the GPLed OS kernel
without your driver linked in, 2) sources to the GPLed OS, 3) a binary
image of your driver as a loadable driver, 4) sources to the
OS-dependent part, and 5) no sources to the OS-independent part. The
situation becomes a little more interesting if this CDROM were to
contain a batch file to *automatically* load the partially-secret
driver into the GPLed OS kernel. Some people have claimed that this
constitutes a de-facto distribution the combined, linked image; I do
not subscribe to that interpretation, as it violates causality. :-)
Referring to Linux specifically, recent kernels have changed
the device driver initiallization process. I believe that these
changes increase the similarity between Linux device drivers and those
of other protected-mode POSIX-like OSes. This should simplify the
task of creating a device driver that is split into OS-dependent and
OS-independent parts, and developing the OS-dependent part for Linux.
Now, in terms of actual practice... if I were facing a market
choice between buying a product with a two-part driver (half of which
were secret, as described above), and buying a product with a
single-part driver (entirely available in free source code), and all
other things were equal... I'd choose the free-source-code product, of
course! If other things weren't equal... I'd still want to choose the
free-source-code product, but I'd try to estimate the economic
tradeoffs first.
In other words, my advice is: know that you can do it, but
please refrain from doing so. :-)
...
>Dynamic linking is just a special case but object oriented development
>is a more general case where derived work don't actually use a single
>line of code from libraries but use them. Why bother to write LGPL if
>anybody may make works derived from GPL'd code without distributing
>sources? If the legal interpretation of GPL allows such behaviours there
>is no need for LGPL at all.
I've argued this position (GPL implies LGPL) in the past.
Again, it is critical to remember that copyright law adresses
"inclusion" (and "translation"), but not "use" (in the non-textual
sense). And again, "conceptual derivation" and "textual derivation"
are two separate matters.
>I don't know what does the GPL or Copyright law consider a derived work.
Then this is where you should concentrate your studies, as it
is essential to have a firm concept of this distinction before
embarking upon debate.
>Could anyone explain wheather the GPL would allow proprietary classes
>derived from a GPL'd class? Does dynamic linking make difference?
Maybe, and yes. There are already court decisions that imply
that the critical content in header files which allows class
derivation to be performed is not subject to Copyright restrictions.
The essential philosophy is that you cannot copyright a verifiable
fact (such as telephone number, but presumably also an offset into a
data structure), only the creative expression in which you present the
fact. However, this is a rather subtle distinction, and court
precedents and legislative action could easily change the situation.
Dynamic linking certainly makes a difference, since static
linking incorporates textual material from multiple sources and
translates it, clearly raising grounds for Copyright infringement,
while dynamic linking attempts to limit textual inclusions to
interface definitions, which may-or-may-not create a Copyright
problem.
You cannot copywright a fact. You can't patent facts, either.
However, you can patent applications of a fact, at least, in the US.
This appears to be a legal basis for using a software patent to
protect an interface, and, hence a logical but non-textual derivation.
Craig Milo Rogers
Terry Lambert
] |> Consider the library function "fooinit":
[ ... ]
] In this case, I would call it a bug. To me it is the same as you
] change the API/ABI. When you do that, the old programs may fail
] no matter if static or dynamic link, nor the source code is provided
] or not. You have to change something to get the old programs working
] again with the new library.
No. I could relink the program objects vs. the new LGPL'ed library,
and it would continue to work.
The problem is that in the prelinked data case, the data/code
abstration for the codified interface has not changed.
We are, in fact, not discussing a library interface change at
all. Unless you are definiing the data provided by the library
and referenced only by the library as part of the interface.
In which case dynamic linking still fails the relink test in LGPL.
This is all real ancillary to the discussion of kernel modules
being loaded as a unit vs. their being dynamically loaded, since
both Linus and I have made clear that the technique in both the
Linux and BSD cases is to load it as a unit, data and all.
I'm quite content to wait for the shared library/relink issue to
be clarified by the FSF (since the LGPL never specifically
mentions data vs. code linking and takes no explicit notice of
shared library technology whatsoever).
Regards,
Michael I. Bushnell, p/BSG
Put it in your licence, if you want. No one forces you to use the
GPL as is. The text of the GPL is public domain, if I recall.
Actually, it's not. It's copyright by the FSF and can only be
distributed if you don't modify it. This is essential to the ability
to use it as a "subroutine". This is the only way to make sure there
is only one thing called the "GNU General Public License".
I think that if you modify it, but clearly call it something else,
there is no problem. But I'm not sure exactly what the rules are.
Michael
John S. Dyson
of GPL. The people that use it often don't understand it very well. This
is not meant as an insult, but this thread scares me further from GPL.
As the proprietor of a small company, this increases my legal expense/exposure.
The license itself is so complex, that the net-wisdom does not seem to converge.
I guess that I'll read it again and try not only to understand the words
said, but any second-order ramifications.
John Dyson
Jimen Ching
>In article <MIB.95De...@gnu.ai.mit.edu>,
>Michael I. Bushnell, p/BSG <m...@gnu.ai.mit.edu> wrote:
>>In article <DJrxt...@kithrup.com> m...@kithrup.com (Mike Stump) writes:
>> Put it in your licence, if you want. No one forces you to use the
>> GPL as is. The text of the GPL is public domain, if I recall.
>>there is no problem. But I'm not sure exactly what the rules are.
>It is *exactly* this kind of statement that makes me even more fearful
>of GPL. The people that use it often don't understand it very well.
Or maybe you do not understand the GPL at all? I've read almost all
permutation of this GPL debate. The only conclusion I can make is that
the people who say the GPL is against commercializing software, does not
understand what the GPL *means* at all.
You say you *read* what the GPL *says*. Fine, but do you understand
what the GPL means? Do you understand the concept of 'software as a
product' vs. 'software as a service'? This is what the GPL *means*.
Through all of the threads that I have read, not one single person
used these two concepts. You say GPL does not allow *commercial use*
of GPL'ed software. Exactly what does that mean? GPL *says* that
source code must be available! How does that affect commercial use?
Do you mean that Walnut Creek can not sell their CDROM's? Or is that
not part of your definition of *commercial use*?
If you don't know by now. What I am implying in this article is that
there is no such thing as *commercial use*. Software is not a product
to be sold or traded. Software is like the procedure for 'heart bypass
surgery'. Once you learned it, you should be able to use it for the
rest of your life. Just like it doesn't make sense to sell the procedure
for 'heart bypass surgery', it doesn't make sense to sell source code.
So what is sold then? What is sold is the skills of the surgeon. Just
like what Cygnus is selling is not software, but the skills of its
software engineers.
This is what the GPL is all about. If you do not believe in this, then
GPL is not for you. But anyone who believe in this philosophy, doesn't
need to understand what the GPL *says*. All they need to know is that
the GPL is what they need to make the above work!
--jc
P.S. Maybe people should stop reading the GPL. They should instead
read the Manifesto. If they believe in the concepts in the Manifesto,
then they should use the GPL. But under no circumstance should anyone
read the GPL. Especially if they're not a lawyer.
--
Jimen Ching (WH6BRR) jch...@aloha.com wh6...@uhm.ampr.org
John S. Dyson
>John S. Dyson (ro...@dyson.iquest.net) wrote:
>>In article <MIB.95De...@gnu.ai.mit.edu>,
>>Michael I. Bushnell, p/BSG <m...@gnu.ai.mit.edu> wrote:
>>>In article <DJrxt...@kithrup.com> m...@kithrup.com (Mike Stump) writes:
>>> Put it in your licence, if you want. No one forces you to use the
>>> GPL as is. The text of the GPL is public domain, if I recall.
>>>I think that if you modify it, but clearly call it something else,
>>>there is no problem. But I'm not sure exactly what the rules are.
>>It is *exactly* this kind of statement that makes me even more fearful
>>of GPL. The people that use it often don't understand it very well.
>
>Or maybe you do not understand the GPL at all? I've read almost all
>permutation of this GPL debate. The only conclusion I can make is that
>the people who say the GPL is against commercializing software, does not
>understand what the GPL *means* at all.
>
(The users of the software, as I DO NOT DEVELOP GPL'ed software) have
the right to use it.
>
>of GPL'ed software. Exactly what does that mean? GPL *says* that
>source code must be available! How does that affect commercial use?
>Do you mean that Walnut Creek can not sell their CDROM's? Or is that
>not part of your definition of *commercial use*?
>
You obviously misunderstand -- you are taking one example of commercial
use and inferring that all other types of use are valid according to the license
and they are not. Your example is simply fallacious.
>
>This is what the GPL is all about. If you do not believe in this, then
>GPL is not for you. But anyone who believe in this philosophy, doesn't
>need to understand what the GPL *says*. All they need to know is that
>the GPL is what they need to make the above work!
>
>P.S. Maybe people should stop reading the GPL. They should instead
>read the Manifesto. If they believe in the concepts in the Manifesto,
>then they should use the GPL. But under no circumstance should anyone
>read the GPL. Especially if they're not a lawyer.
>
Oh my gosh!!! REMEMBER, the GPL is the license that YOU MUST OBEY to
use the software. Very little outside of the license is important (as you risk
legal action if you do not abide by what the GPL says.) I for one am
not willing to be dependent on the whims of some software developer
somewhere when he *thinks* that I have violated his license, even though
I *feel* that I understand other supporting documents.
I guess I just won't read the license and be all touchy feely and trust
that I am one with the one true GPL god :-). That is NOT the way to
do business... It is apparent to me that there is a community of people
out there who do not understand the risks taken by not reading contracts
carefully. And after reading them, I am convinced that these people
COULD not understand the side-effects of the agreements that they are taking
on. It is a sorry state of affairs...
The only valid discussion that I can see on this subject is to discuss the
terms and requirements of the GPL. It is also valid to compare GPL vs. BSD
vs. Artistic, ad nauseum... Other supporting documents are interesting
BUT NOT THE AGREEMENT(S) THAT YOU MUST ADHERE TO...
John Dyson
Paul D. Smith
%% ro...@dyson.iquest.net (John S. Dyson) writes:
jsd> In article <MIB.95De...@gnu.ai.mit.edu>,
jsd> Michael I. Bushnell, p/BSG <m...@gnu.ai.mit.edu> wrote:
>> Actually, it's not. It's copyright by the FSF and can only be
>> distributed if you don't modify it. This is essential to the
>> ability to use it as a "subroutine". This is the only way to
>> make sure there is only one thing called the "GNU General Public
>> License".
>>
>> I think that if you modify it, but clearly call it something
>> else, there is no problem. But I'm not sure exactly what the
>> rules are.
jsd> It is *exactly* this kind of statement that makes me even more
jsd> fearful of GPL. The people that use it often don't understand
jsd> it very well. This is not meant as an insult, but this thread
jsd> scares me further from GPL. As the proprietor of a small
jsd> company, this increases my legal expense/exposure.
Again, I don't understand this position. Suppose you took the licensing
scheme from your Microsoft program, labelled "Microsoft's License",
modified it slightly, then used it with your program still calling it
"Microsoft's License". That's obviously not a good idea, right? So why
the amazement that the GPL is the same way? Where's the extra exposure
there?
As for people using the license not understanding it, again, where's the
big difference between the GPL and commercial software licenses? I bet
if you asked 100 net denizens detailed questions about the licensing
that came with their last piece of commercial software at least 50 would
get significant portions wrong, and advise you incorrectly.
Heck, from what I can see on many newsgroups, more than 50% of the
people posting don't have any idea what they're talking about on _any_
given subject :)
jsd> The license itself is so complex, that the net-wisdom does not
jsd> seem to converge.
I maintain that it's not the _license_, at *all*, that's complex. It's
the obscure situations that people come up with on this newsgroup (and
elsewhere), either to exercise the boundaries of the license for their
own intellectual edification, or because they're trying to find a way to
"cheat" the license. If this newsgroup spent as much time trying to
examine the details of commercial licenses, I daresay you'd be even more
nervous about them than the GPL! :)
If you use the GPL in a straightforward manner, there's no mystery
there. If you think there is, or are worried, describe the specific
situation and I'm sure people will point you to the parts of the GPL
they feel relate to your situation.
Then, as with any license, it's up to you to read and comprehend, and
finally to decide to take the risk or not. The risk may be real, but I
can't see that it's greater with the GPL than any other license. I
personally feel it's less.
-------------------------------------------------------------------------------
Paul D. Smith <psm...@baynetworks.com> Network Management Development
Senior Software Engineer Bay Networks, Inc.
-----------------------------------------------==<http://www.baynetworks.com/>-
"Please remain calm...I may be mad, but I am a professional." --Mad Scientist
-------------------------------------------------------------------------------
These are my opinions--Bay Networks takes no responsibility for them.
John S. Dyson
Paul D. Smith <psm...@baynetworks.com> wrote:
>%% Regarding Re: [Q] GPL violation.;
>%% ro...@dyson.iquest.net (John S. Dyson) writes:
>
> jsd> It is *exactly* this kind of statement that makes me even more
> jsd> fearful of GPL. The people that use it often don't understand
> jsd> it very well. This is not meant as an insult, but this thread
> jsd> scares me further from GPL. As the proprietor of a small
> jsd> company, this increases my legal expense/exposure.
>
>Again, I don't understand this position. Suppose you took the licensing
>scheme from your Microsoft program, labelled "Microsoft's License",
>modified it slightly, then used it with your program still calling it
>"Microsoft's License". That's obviously not a good idea, right? So why
>the amazement that the GPL is the same way? Where's the extra exposure
>there?
>
been trying to read something into the GPL that isn't there I guess...
>
>As for people using the license not understanding it, again, where's the
>big difference between the GPL and commercial software licenses?
>
If there is no difference to me, then what makes GPL so good???
There are so many interpretations of what GPL means, and a person in a
previous posting made a very constructive point... It would be good to
find a lawyer that is not connected with FSF at all to analyze the GPL
and write an unbiased laymans' version of it.
My view of it is that GPLed code isn't alot better than Microsoft's code
when it comes to me making money. I mean -- I can make money supporting
Microsoft OSes, Compilers, etc... We have such a support organization where
I work for example. So can I supporting GPLed code... But if I try to resell
portions of Microsoft code, I can get in trouble, and likewise GPLed code...
In fact, it is apparently fairly easy to violate the GPL, as it is to violate
the license for Microsoft C.
John Dyson
Mike Stump
>The people that use it often don't understand it very well.
I always thought the central concept of the GPL was rather simple
myself.
>This is not meant as an insult, but this thread scares me further
>from GPL. As the proprietor of a small company, this increases my
>legal expense/exposure.
One can dump as much money as they care into legal. If you are GPLing
your own code, the expense/exposure can be low, very low.
If you are using other's GPLed code, you can always run the use by
them, and see if they balk. If they don't, then you fairly ok from a
pratical perspective (if you trust them). If they do, then it doesn't
matter what your lawyer thinks as they will want to take you to court
anyway.
There is a cost to using GPLed code, yes. For most uses, the cost is
kept down to a reasonable level. The higher the value of the use of
the GPLed code, the more cost is in using it. This isn't unreasonable.
For example, to use emacs to edit files, it is a pretty good no
brainer. But, if you want to edit it up, and sell it as the basis for
your own C++ code browser, better dump some money into it.
>The license itself is so complex,
It just seems that way. :-)
>that the net-wisdom does not seem to converge.
I disagree. I have only ever seen one unresolved disagreement before.
And for that case, one should consult his lawyer, and measure the
risk/benefit.
If you feel there are any unresolved points, state them here, and then
just sit back, and watch the discussion. In the end, I think it
should be clear.
Mike Stump
>the Manifesto.
I disagree.
For the newcomers to this list, please don't read the Manifesto,
decide you disagree in part, or in all with it, and then assume that
the GPL isn't for you.
I disagree, in part at least, with the Manifesto, but I don't hold
that against the GPL, nor rms, nor the FSF, nor emacs.
John S. Dyson
>In article <4bgt6i$j...@dyson.iquest.net>,
>John S. Dyson <ro...@dyson.iquest.net> wrote:
>
>One can dump as much money as they care into legal. If you are GPLing
>your own code, the expense/exposure can be low, very low.
>
of GPL is interesting from the standpoint of the original author loosing
ownership of his/her software as mods and enhancements are added by the
net at large. Eventually, ownership can become so diffuse, that licensing
for commercial use can be tricky. There are licenses that totally eliminate
that problem. (I can take code that I originally wrote, but others have
enhanced, and perhaps put it into an embedded controller that I just do not
want to deal with source redistribution on, for example.) There is
no sinister (sorry, lefties) plan here...
>
>There is a cost to using GPLed code, yes. For most uses, the cost is
>kept down to a reasonable level. The higher the value of the use of
>the GPLed code, the more cost is in using it. This isn't unreasonable.
>
I am pretty much forbidden from using GPLed code at work (in product), but
of course can use it in development tools as I would purchased software
from Microsoft. I almost definitely would have to license the GPLed
work seperately, and that is not even practical in some cases. We are
NOT about to redistribute source code -- sorry, just doesn't/will not
happen.
>
>If you feel there are any unresolved points, state them here, and then
>just sit back, and watch the discussion. In the end, I think it
>should be clear.
>
Actually, my questions have been an attempt at leading the discussion
into whether GPLed software is as encumbered as to redistribution as it
appears that it is. This IS NOT MEANT TO BE FLAMEBAIT, but as an example,
the Berkeley Networking Code is used as a seed for many TCP/IP suites
(for better or worse.) This would not be practical under GPL without
getting seperate agreement from the authors. The notion of GPLed software
as being a community resource is diminished by this.
Tell me where I am wrong...
(These are a few of my worries, more to come.)
John Dyson
Russ Allbery
> Since I am a substantial contributor to a free software project, the issue
> of GPL is interesting from the standpoint of the original author loosing
> ownership of his/her software as mods and enhancements are added by the
> net at large. Eventually, ownership can become so diffuse, that licensing
> for commercial use can be tricky. There are licenses that totally
> eliminate that problem. (I can take code that I originally wrote, but
> others have enhanced, and perhaps put it into an embedded controller that
> I just do not want to deal with source redistribution on, for example.)
> There is no sinister (sorry, lefties) plan here...
Let me get this straight...your objection to the GPL is that it doesn't
automatically give you the right to distribute code someone else wrote? Of
course it doesn't. I'd be interested to see the license that does, and
whether it would hold up in court.
Your *original* code, along with any modifications you write yourself, is
your possession. You can release it under any license you care to use. You
do not automatically get the right to release other people's improvements
under a different license without making additional arrangements with them.
If you think this is a problem unique to the GPL, I believe you're wrong.
I'm sure Sun (for example) needs to be just as meticulous about maintaining
ownership (or ensuring that the license agreements are compatible with their
use) of all the code that goes into Solaris. They're undoubtably aided by
employee contracts with the people working on the code.
If they were to open Solaris up to modification by the Net at large, I'm
sure they would have to make legal arrangements with the authors of
improvements before they could incorporate them into Solaris under the
standard licensing agreement.
(The example of Solaris is used solely because it was the first thing that
came to mind. If I am mistaken about Sun's licensing scheme, please
substitute some other commercially licensed, non-GPL product (Windows '95,
for instance).)
> I am pretty much forbidden from using GPLed code at work (in product), but
> of course can use it in development tools as I would purchased software
> from Microsoft. I almost definitely would have to license the GPLed work
> seperately, and that is not even practical in some cases. We are NOT
> about to redistribute source code -- sorry, just doesn't/will not happen.
Yes, in that case, you're not going to be able to use GPL code in your
product without making additional arrangements with the copyright
holder(s). I'm not sure why you consider this a criticism of the GPL.
There *is* a political agenda behind the GPL -- it's stated in the
Manifesto. I would imagine that most of us who release code under GPL agree
with that political agenda, at least to some extent (if only to the extent
that we feel the user of the software should have the freedom to modify it
as they wish). By not releasing the source code of your product, you are
acting directly contrary to that political agenda, and therefore the GPL is
going to make matters difficult for you.
But note that the GPL only applies to code that you have received *under
that license*. It doesn't apply to code you wrote yourself, since you have
ownership of that code and can ignore any licensing agreement. It doesn't
apply to code released under the GPL but also released under another
licensing agreement, which you are using under that other licensing
agreement (Perl under the Artistic License, for instance). It only applies
to code that you would not have been able to use were it not under GPL.
In no way can the GPL cause the owner of the code to lose possession of it.
> This IS NOT MEANT TO BE FLAMEBAIT, but as an example, the Berkeley
> Networking Code is used as a seed for many TCP/IP suites (for better or
> worse.) This would not be practical under GPL without getting seperate
> agreement from the authors. The notion of GPLed software as being a
> community resource is diminished by this.
It may be in your opinion. It isn't in mine, because I believe in source
code distribution and *want* to have the source code for software packages
that I use and *want* to distribute the source code of products I write.
The community of people who are benefited by GPLed software are those who
release their source code. There are some benefits to those who do not, but
I personally believe those benefits are mainly nice side effects, not the
main purpose of the license.
I'm sure that I'll be corrected by others here who know more of the
specifics of the situation than I....
--
Russ Allbery (r...@cs.stanford.edu) http://www-leland.stanford.edu/~rra/
Russ Allbery
>> Let me get this straight...your objection to the GPL is that it doesn't
>> automatically give you the right to distribute code someone else wrote?
>> Of course it doesn't. I'd be interested to see the license that does,
>> and whether it would hold up in court.
> The BSD license does. With a very simple set if stipulations... Even
> USL dealt with it unhappily (existance proof.)
I'd appreciate more details on this, because it certainly doesn't look like
it does. All the BSD license does is list provisions under which your
program can be distributed. Nowhere does it say "any modifications to this
program become the property of the original author" or "any modifications to
this program can be distributed under the BSD license."
I could take your BSD licensed program, make my own modifications to it, and
then distribute my modifications (in some encapsulated form not containing
the original source) under a commerical license and you wouldn't be able to
use them in your product without working it out with me.
>> Your *original* code, along with any modifications you write yourself, is
>> your possession. You can release it under any license you care to use.
>> You do not automatically get the right to release other people's
>> improvements under a different license without making additional
>> arrangements with them.
> Sorry, then it does not get the advantage of enhancements made over time
> on the net... Bzzzt... (Sort-of a trade for free-use on the net for bug
> fixes and enhancements that can be used as opposed to encumbered with
> GPL.)
This has nothing to do with the GPL. You're *always* going to have to make
arrangements with the people who produced the code modifications you want to
use if you want to incorporate them in a commerical product unless they
explicitly made them public domain or put them under a license that allowed
you to do what you want. This is going to be true regardless of whether you
distribute your code under the GPL or not.
> The notion of GPLed software being freeware appears to be double speak
> that is clearly and blatently set forth in the GNU Manifesto...
Where on earth did you get the impression that GPLed software was freeware?
Didn't you ever read the Manifesto? To quote:
GNU is not in the public domain. Everyone will be permitted to
modify and redistribute GNU, but no distributor will be allowed to
restrict its further redistribution. That is to say, proprietary
modifications will not be allowed. I want to make sure that all
versions of GNU remain free.
GPLed software is *free*. It is not freeware.
> Ok, but how does that relate to true freeware???... Certain copyrights
> only require attribution.
If you prefer that, use it. I'm not particularly interested in what you
call "true freeware."
> I guess I don't subscribe to that agenda, and I think that people should
> read it VERY CAREFULLY with understanding of it's side effects before
> signing up... :-). The agenda is not only political but has the effect of
> converting "ownership" from the original author to the net at large or
> large support entities. This conversion process takes time and appears to
> be insidious. (Defining "ownership" as the ability to profit in one way
> or another.)
What a bizarre definition of ownership. If by profit, you mean make money,
then apparently I don't own my couch. If by profit, you mean derive
benefts from, then everyone owns GPLed software.
I have to say I don't find your definition of ownership particularly useful.
John S. Dyson
Russ Allbery <r...@cs.stanford.edu> wrote:
>John S Dyson <ro...@dyson.iquest.net> writes:
>
>> Since I am a substantial contributor to a free software project, the issue
>> of GPL is interesting from the standpoint of the original author loosing
>> ownership of his/her software as mods and enhancements are added by the
>> net at large. Eventually, ownership can become so diffuse, that licensing
>> for commercial use can be tricky. There are licenses that totally
>> eliminate that problem. (I can take code that I originally wrote, but
>> others have enhanced, and perhaps put it into an embedded controller that
>> I just do not want to deal with source redistribution on, for example.)
>> There is no sinister (sorry, lefties) plan here...
>
>Let me get this straight...your objection to the GPL is that it doesn't
>automatically give you the right to distribute code someone else wrote? Of
>course it doesn't. I'd be interested to see the license that does, and
>whether it would hold up in court.
>
USL dealt with it unhappily (existance proof.)
>
>Your *original* code, along with any modifications you write yourself, is
>your possession. You can release it under any license you care to use. You
>do not automatically get the right to release other people's improvements
>under a different license without making additional arrangements with them.
>
Sorry, then it does not get the advantage of enhancements made over time
on the net... Bzzzt... (Sort-of a trade for free-use on the net for bug
fixes and enhancements that can be used as opposed to encumbered with GPL.)
>
>If you think this is a problem unique to the GPL, I believe you're wrong.
>I'm sure Sun (for example) needs to be just as meticulous about maintaining
>ownership (or ensuring that the license agreements are compatible with their
>use) of all the code that goes into Solaris. They're undoubtably aided by
>employee contracts with the people working on the code.
>
Sure, and Sun does not purport to give away it's freeware (SunOS/Solaris) :-).
The notion of GPLed software being freeware appears to be double speak that
is clearly and blatently set forth in the GNU Manifesto...
>
>If they were to open Solaris up to modification by the Net at large, I'm
>sure they would have to make legal arrangements with the authors of
>improvements before they could incorporate them into Solaris under the
>standard licensing agreement.
>
Ok, but how does that relate to true freeware???... Certain copyrights
only require attribution.
>
>There *is* a political agenda behind the GPL -- it's stated in the
>Manifesto. I would imagine that most of us who release code under GPL agree
>with that political agenda, at least to some extent (if only to the extent
>that we feel the user of the software should have the freedom to modify it
>as they wish). By not releasing the source code of your product, you are
>acting directly contrary to that political agenda, and therefore the GPL is
>going to make matters difficult for you.
>
I guess I don't subscribe to that agenda, and I think that people should
read it VERY CAREFULLY with understanding of it's side effects before
signing up... :-). The agenda is not only political but has the effect of
converting "ownership" from the original author to the net at large or
large support entities. This conversion process takes time and appears
to be insidious. (Defining "ownership" as the ability to profit in one
way or another.)
John Dyson
Jimen Ching
>For the newcomers to this list, please don't read the Manifesto,
>decide you disagree in part, or in all with it, and then assume that
>the GPL isn't for you.
I read both the Manifesto and the GPL. I've always thought the Manifesto
is the 'why' and the GPL is the 'how' of free[dom] software. If you
understand the 'why', then the 'how' isn't that important, especially if
you are not into the legal details.
>I disagree, in part at least, with the Manifesto, but I don't hold
>that against the GPL, nor rms, nor the FSF, nor emacs.
I might be able to get a better understanding of your point of view
if you can provide passages in the Manifesto which would direct people
against the GPL.
--jc
Jimen Ching
>large support entities. This conversion process takes time and appears
>to be insidious. (Defining "ownership" as the ability to profit in one
>way or another.)
I do not understand this. In practically every article you posted, you
say in one way or another, that GPL prevents the author from making a
profit. You mean by selling the source code? No one prevents you from
selling source code. Walnut Creek sell multiple CDROM's of GPL code.
You can do the same with GPL code. What other profit do you have in
mind? The following are some other ways to earn a profit from giving
source code away.
Crynwyr drivers
Cygnus Support
Remember, the point of the GPL is to provide source code. So if you
suggest selling 'software as a product' like Netscape, you've already
misunderstood the purpose of GPL. The GPL is specifically designed to
fight against this. Are you criticizing GPL for doing what it was
designed to do?
Russ Allbery
>> I could take your BSD licensed program, make my own modifications to it, and
>> then distribute my modifications (in some encapsulated form not containing
>> the original source) under a commerical license and you wouldn't be able to
>> use them in your product without working it out with me.
> Cool... That is the purpose of my gift... There is no commercial intent,
> but any free versions are still fair game. At least in the BSD community
> the free versions are keeping up with the commercial versions... I am not
> afraid of the commercial stuff. I am more afraid of encumbered code in
> the public where I can see the code and I have to clean-room reimplement
> it.
Then I definitely didn't understand your original point. You don't seem to
be concerned with what license you yourself use, but rather what license the
people making modifications to your programs use.
Maybe I can explain how this sounds to me. You want to maintain your right
to release commercial versions of your software, but you want all
modifications to your software to be released under a license that lets you
make money off of them.
This immediately turns me off.
Now I'm not sure you actually feel that way, since you've said that you're
part of major free software products. I'm not sure where you stand on
this. But to me, this feels liking asking for Christmas gifts so that you
can sell them. My instinctive response to the offering of free software is
to respond in kind and release my own free software, not to say "hey, thanks
for the software" and then going off to sell it.
Now don't get me wrong; you should certainly be able to sell your
programming experience and the products of your labors. What I highly
dislike is the idea of selling *other people's* labors without paying them
for it. I realize that the GPL also allows this, but it at least requires
that source code also be developed so that others can distribute the work
for free.
> All I want is a little bit of credit if someone uses my code -- that is
> all.
Then perhaps you should use the BSD license. The reason why *I* would use
the GPL is because it encourages the development of more free software,
whereas the BSD license does not.
>> GPLed software is *free*. It is not freeware.
> The old double-speak again. GPL software is NOT free software to me.
The users of GPL software are free to modify it and improve it. They are
free to distribute their versions for other people to use. They are free to
use the code in their own packages provided that their packages are also
free.
This is free software to me. Apparently you have a different definition,
and that's fine. Most words have different definitions depending on who you
ask. But accusing someone of doublespeak because you disagree with their
definitions is dishonest.
> Cygnus and other support companies are free to make money from it -- so
> therefore it is free. :-). But since I write code for companies that
> believe that proprietary is the way to go (which in my commercial work --
> I believe also), I am not free to use it...
Yes, your *company* is not free to use it in any way they please because
your company is not part of the free software community.
> I CAN use it in development -- and I do, up to the limits of the GPL. But
> the GPL limits me -- sorry, it is just broken for me in much of my work.
It's not broken; it's doing precisely what it was designed to do. (Be aware
that my political opinions are not necessarily those of other people who use
the GPL.)
>> If you prefer that, use it. I'm not particularly interested in what you
>> call "true freeware."
> My interests are apparently not the same as yours... I hope that is what
> you mean, and are not just blowing me off -- that is impolite...
No, you have the intention correct. I would be perfectly happy to see the
commerical software market shrivel up and largely die. Writing software
that is free to be used in commerical software packages without source
distribution is not high on my priority list.
I'm paid for *writing* software, not for *having written* software. Other
users of the GPL probably have different perspectives.
John S. Dyson
>
>>> Let me get this straight...your objection to the GPL is that it doesn't
>>> automatically give you the right to distribute code someone else wrote?
>>> Of course it doesn't. I'd be interested to see the license that does,
>>> and whether it would hold up in court.
>
>> USL dealt with it unhappily (existance proof.)
>
>it does. All the BSD license does is list provisions under which your
>program can be distributed. Nowhere does it say "any modifications to this
>program become the property of the original author" or "any modifications to
>this program can be distributed under the BSD license."
>
>then distribute my modifications (in some encapsulated form not containing
>the original source) under a commerical license and you wouldn't be able to
>use them in your product without working it out with me.
>
Cool... That is the purpose of my gift... There is no commercial intent, but
any free versions are still fair game. At least in the BSD community the
free versions are keeping up with the commercial versions... I am not
afraid of the commercial stuff. I am more afraid of encumbered code in the
my area of work on free software, there aren't many organizations that can do
much better -- certainly not afraid of them. I AM afraid of tainted code in
the public. All I want is a little bit of credit if someone uses my code --
that is all.
>
>This has nothing to do with the GPL. You're *always* going to have to make
>arrangements with the people who produced the code modifications you want to
>use if you want to incorporate them in a commerical product unless they
>explicitly made them public domain or put them under a license that allowed
>you to do what you want. This is going to be true regardless of whether you
>distribute your code under the GPL or not.
>
That is why I like the BSD license. Again, I find that the free software
projects tend to lead the commercial entries. -- I am NOT afraid of them.
In fact, I aplaud the ability of a small start-up being able to adopt code
and make money from it. They don't even have to give the results of their
work to their competition!!!
>
>Where on earth did you get the impression that GPLed software was freeware?
>Didn't you ever read the Manifesto? To quote:
>
> GNU is not in the public domain. Everyone will be permitted to
> modify and redistribute GNU, but no distributor will be allowed to
> restrict its further redistribution. That is to say, proprietary
> modifications will not be allowed. I want to make sure that all
> versions of GNU remain free.
>
>GPLed software is *free*. It is not freeware.
>
The old double-speak again. GPL software is NOT free software to me. Cygnus
and other support companies are free to make money from it -- so therefore it
is free. :-). But since I write code for companies that believe that
proprietary is the way to go (which in my commercial work -- I believe also),
I am not free to use it... I CAN use it in development -- and I do, up
to the limits of the GPL. But the GPL limits me -- sorry, it is just
broken for me in much of my work.
>
>If you prefer that, use it. I'm not particularly interested in what you
>call "true freeware."
>
My interests are apparently not the same as yours... I hope that is what
you mean, and are not just blowing me off -- that is impolite...
>
>I have to say I don't find your definition of ownership particularly useful.
>
I meant in the sense as a tool in my work. If I don't have free use of
it I cannot profit from it. Sorry for your misunderstanding of what
I have been trying to say. I have been answering lots of email privately
on this subject also -- just tired, sorry ...
For example:
Cygnus is in a interesting postion of being able to profit from GPLed software
that is perhaps written by other people. I can see where GPL is good for it,
but not me... I could be convinced that GPL is good for me if I can
benefit from it -- it is obvious that Cygnus is convinced that GPL is good
for it.
John Dyson
Gasparovski / Daniel (ISE)
>
> Put it in your licence, if you want. No one forces you to use the
> GPL as is. The text of the GPL is public domain, if I recall.
>
>distributed if you don't modify it. This is essential to the ability
>to use it as a "subroutine". This is the only way to make sure there
>is only one thing called the "GNU General Public License".
>
>I think that if you modify it, but clearly call it something else,
>there is no problem. But I'm not sure exactly what the rules are.
Maybe the FSF should release the GNU General Public License under a
BSD-like license. You know, you can copy it and what not but if you
change it don't call it the official GNU GPL (as per clause 4). :)
Dan ...
--
[ Danny Gasparovski | Mortified by the lack of conscience ]
[ u92...@student.canberra.edu.au | Our sanctity bears no relevance ]
[ University of Canberra, Australia | Insignificant is our existence ]
[ Bolt Thrower, "The IVth Crusade" -> | Hear the litany of life's persistence ]
Ingemar Hulthage
Russ Allbery <r...@cs.stanford.edu> first wrote:
> How am I imposing my will on anyone? If you don't like my terms,
> don't use my code. That's no different than any other licensing
> arrangement. The BSD license is forcing you to publicize the names
> of the original authors whether you like it or not.
later he writes:
> The reason why *I* would use the GPL is because it encourages the
> development of more free software, whereas the BSD license does not.
Where does the line between 'imposing' and 'encouraging' go ? If
there is no coercing going on, why not make GPL voluntary also for
derivative work ?
Also there are important differences between the GPL and 'other
licensing arrangements'. For example, when a copyright holder
reserves the rights, associated with copyright, in order to get
economical compensation or acknowledgement, he is seeking legitimate
interests. But when the GPL coerces people to use the GPL for
derivative work, no legitimate interest of the copyright holder is
protected. The copyright is only used as a leverage to make other
people conform to certain values. A copyright holder may have the
right to use the GPL license for his code, but it smacks of
intolerance and extremism.
Merry Christmas,
Ingemar Hulthage
(This message is my copyright and may not be quoted unless the phrase
'Merry Christmas' is included 8-^)
William S. Gribble
> Also there are important differences between the GPL and 'other
> licensing arrangements'. For example, when a copyright holder
> reserves the rights, associated with copyright, in order to get
> economical compensation or acknowledgement, he is seeking legitimate
> interests. But when the GPL coerces people to use the GPL for
> derivative work, no legitimate interest of the copyright holder is
> protected.
I would say that your definition of ``legitimate interest'' is far too
narrow. The GPL serves the interest of the copyright holder by making
the world of free software just a little bit richer, by making software
in general just a little bit easier to get, use, and improve, and
by fractionally reducing the stranglehold that proprietary software
has on the marketplace. I'd say tht's definitely a legitimate interest
of a programmer.
> A copyright holder may have the right to use the GPL license for his
> code, but it smacks of intolerance and extremism.
.. and that statement smacks of hyperbole and bombast.
> Merry Christmas,
> (This message is my copyright and may not be quoted unless the phrase
> 'Merry Christmas' is included 8-^)
Actually, that's not enforceable.. quoting is fair use. Not to mention
that I hate Christmas, and it's already over :) But in the interest
of altruism I'll reproduce your message.
Bill Gribble
Barry Margolin
Ingemar Hulthage <hult...@hollywood.cinenet.net> wrote:
>Russ Allbery <r...@cs.stanford.edu> first wrote:
>> development of more free software, whereas the BSD license does not.
>Where does the line between 'imposing' and 'encouraging' go ? If
>there is no coercing going on, why not make GPL voluntary also for
>derivative work ?
The difference between "imposing" and "encouraging" is that you aren't
forced to write derivative works of GPLed code. You have a choice: you can
write proprietary code, or you can build upon the existing GPLed code
base. The existence of a high-quality base of free code is used as an
incentive to get you to create more code like it.
--
Barry Margolin
BBN PlaNET Corporation, Cambridge, MA
bar...@bbnplanet.com
Phone (617) 873-3126 - Fax (617) 873-6351
Russ Allbery
> Russ Allbery <r...@cs.stanford.edu> first wrote:
>> How am I imposing my will on anyone? If you don't like my terms,
>> don't use my code. That's no different than any other licensing
>> arrangement.
This is misphrased, because it implies that I'm denying that I'm imposing my
will. I'm not; I'm saying that I'm imposing my will only on my code and
through a licensing agreement, and that the BSD license is doing the same
thing.
> Where does the line between 'imposing' and 'encouraging' go? If there is
> no coercing going on, why not make GPL voluntary also for derivative work?
All licenses coerce.
> Also there are important differences between the GPL and 'other licensing
> arrangements'. For example, when a copyright holder reserves the rights,
> associated with copyright, in order to get economical compensation or
> acknowledgement, he is seeking legitimate interests. But when the GPL
> coerces people to use the GPL for derivative work, no legitimate interest
> of the copyright holder is protected.
I find it convenient how you define legitimate to suit your argument.
Let's just say that I do not accept your definition of legitimate.
> A copyright holder may have the right to use the GPL license for his code,
> but it smacks of intolerance and extremism.
More so than a commerical license? I'd have to say that you're extremely
confused. Someone is intolerant and extreme because they release their code
for the public to modify and use?
Marcus Daniels
In article <HULTHAGE.95...@hollywood.cinenet.net> hult...@hollywood.cinenet.net (Ingemar Hulthage) writes:
Ingemar> The copyright is only used as a leverage to make
Ingemar> other people conform to certain values.
Hey, you're catching on!
Ingemar> A copyright holder
Ingemar> may have the right to use the GPL license for his code, but
Ingemar> it smacks of intolerance and extremism.
Hoarding the code of derived work smacks of exploitation. Go figure.
[and here's hoping that "Merry Christmas" resulted in big bucks all around!]
John S. Dyson
Barry Margolin <bar...@tools.bbnplanet.com> wrote:
>In article <HULTHAGE.95...@hollywood.cinenet.net>,
>Ingemar Hulthage <hult...@hollywood.cinenet.net> wrote:
>>> development of more free software, whereas the BSD license does not.
>
>>there is no coercing going on, why not make GPL voluntary also for
>
>The difference between "imposing" and "encouraging" is that you aren't
>forced to write derivative works of GPLed code. You have a choice: you can
>write proprietary code, or you can build upon the existing GPLed code
>base. The existence of a high-quality base of free code is used as an
>incentive to get you to create more code like it.
Third choice -- you can build upon the BSDed code base... It works also.
There are glitches in the system, but there are in GPL also, they are just
less visible, because there are companies and individuals that
stay away from the GPL license a priori.
John Dyson
John S. Dyson
Marcus Daniels <mar...@sysc.pdx.edu> wrote:
>>>>>> "Ingemar" == Ingemar Hulthage <hult...@hollywood.cinenet.net> writes:
>In article <HULTHAGE.95...@hollywood.cinenet.net> hult...@hollywood.cinenet.net (Ingemar Hulthage) writes:
>
>Ingemar> The copyright is only used as a leverage to make
>Ingemar> other people conform to certain values.
>
>Hey, you're catching on!
>
>Ingemar> A copyright holder
>Ingemar> may have the right to use the GPL license for his code, but
>Ingemar> it smacks of intolerance and extremism.
>
>Hoarding the code of derived work smacks of exploitation. Go figure.
>
I DO NOT feel exploited when someone makes money from what I do. Since R+D
budgets might be about 5% of gross revenues, and I get perhaps 1-10% of the R+D
budget (if I am lucky), am I being exploited? -- NO. I agreed to those
terms and it is reality in the marketplace -- and so the BSD license allows
people to agree to share code if they want to -- but it does not compel them
to. Why should I worry if someone is ingenuous(sp) enough to find a way to make
money on software that is public??? They must have done *significant* work
on the code for the free sector to not be able to reproduce it.
John Dyson
Marcus Daniels
In article <4c3hnp$3...@dyson.iquest.net> ro...@dyson.iquest.net (John S. Dyson) writes:
Ingemar> A copyright holder may have the right to use the GPL license
Ingemar> for his code, but it smacks of intolerance and extremism.
me> Hoarding the code of derived work smacks of exploitation. Go
me> figure.
John> I DO NOT feel exploited when someone makes money from what I do.
Who said anything about money? Hoarding is the activity of taking
something that was *intended* for the public commons and making it
a product for private profit. GNU developers *intend* this, hence
they select the GPL. I don't think you will find many GNU
developers who would categorically object to the notion of their
work resulting in the production of wealth for others.
