license question?

[Floyd L. Davidson]

xarax <[email protected]> wrote:

Again, you are a moron. I've spent years
negotiating license agreements with attorneys,
CEOs, and CFOs.

What the
LGPL does is allow you to distribute such a derived work (and I quote)
"under terms of your choice" [provided only that you allow reverse
engineering and debugging and relinking and access to the source code
of the Library, to summarize your obligations].

Nope. Derivative work imposes restrictions and
demands on *your* code. Read the fucking agreement
again from end to end.

While you were emptying the trash cans and ashtrays in these
negotiating sessions, did anyone ever explain to you that if it
is *not* a derivative work then *no* license is required at all?

Clearly Peter is correct.

And never ever use any GNU-type license for
code that you want retain as your own.

Unless of course you want to retain it, *and* be guaranteed free
access to all derivative works.

 

[Peter]

Well, there's the obvious commercial products from Linux distributors,
but I doubt that's what you meant. Sitecom GmbH in Germany was found
to be using the GPL'd netfilter code from Linux in their wireless
routers and access points. The Buffalo (Melco) WBR-G54 used as the
base for wireless products by Belkin, Buffalo, Linksys and possibly
others was based around Linux, BusyBox and other GPL'd software. You'll
find many operating systems, including Solaris, include large amounts
of GPL'd software for basic system utilities.

Of course, as a downside the most common reason for a company to be
publicly advertised as using GPL'd code is for the owner of the code
to notice them distributing his code but not honoring the terms of the
GPL while doing so. I'll have to dig them up, but I recall several such
cases being settled this year alone. The FSF could probably give you
more information on those.

Oh shit, i don't think Buffalo's product is open source. FSF should
bring them to the court.
from Peter

 

[Steve Schefter]

False. GPL or LGPL is NOT viral. You may do with your work what you wish.
However, the programs licensed under GPL and LGPL has certain terms under
which you are allowed to copy them If you do not abide by those terms then
you are not allowed to use those products. Your work does NOT become GPL.
Your work is whatever you say. HOwever, you may be violating the license of
those whose work you used, and they could sue you if you do not follow
their license.

This has been discussed in other newsgroups and the question of the viral
nature of GPL continues. There are differing opinions.

First, to define what I mean by viral nature: It is unrelated to what
you do about distributing the original GPL source or binaries produced from
it (at least not directly distributing -- see below). Rather, it has to do
with distributing your own code/binaries. It may be the most clear to define
by an example. My software calls the function blahblah() passing a certain
set of parameters that it expects. I learned what parameters to pass by
looking at the source for that function. Given that, if (I say if -- I offer
no opinion here) the license of blahblah() is indeed viral, then my code is
deemed to be a derivative work, arguing that I could not have produced my
work without theirs. It's up to the lawyers to argue if this does constitute
a derivative work and if GPL fits into that or not. Also, this is just an
example. There are other ways in which it may be clained that you "pick up"
into your code the license of the code you didn't write but work with.

This situation is of more interest to us who write in the kernel space rather
than application space since there is less documentation for the kernel
interfaces. In many cases, you only know what to do by reading the source
for those routines (an other one of my beefs, but I'll stay off that soap box).

For what it's worth, there isn't even agreement between Linus and Alan Cox
on this. On the question of whether a driver writer is required to put his
code under GPL simply because it is insmoded with a GPL kernel or if it
can remain closed source (in binary form only) I quote a reply made by
Alan Cox:

And this is legal according to the "Kernel GPL, Linus Torvalds edition
(TM)" which says "any loadable module can be binary only". Not "only
loadable modules which are drivers". It may not be the intention but
it is the fact.

Linus opinion on this is irrelevant. Neither I nor the FSF nor many others
have released code under anything but the vanilla GPL. By merging such code
Linus lost his ability to vary the license.

So it comes down to the question of whether the module is linking (which is
about dependancies and requirements) and what the legal scope is. Which
is a matter for lawyers.

Anyone releasing binary only modules does so having made their own
appropriate risk assessment and having talked (I hope) to their insurers

 

[Shmuel (Seymour J.) Metz]

I would be very careful of doing that. You would certainly want to
be able to link to several alternative versions of the library, and
make it the user's choice (and responsibility) to decide to which
library they would link.

I'd probably give a choice between his currently installed copy of the
GPL material or the copy included in my distribution, along with a
list of mirror sites in case he wants to instal a different version
prior to proceeding. I don't think that I'd want to package more than
one version in my distribution.

--
Shmuel (Seymour J.) Metz, SysProg and JOAT <http://patriot.net/~shmuel>

Unsolicited bulk E-mail subject to legal action. I reserve the
right to publicly post or ridicule any abusive E-mail. Reply to
domain Patriot dot net user shmuel+news to contact me. Do not
reply to (e-mail address removed)

 

[Malcolm Dew-Jones]

Floyd L. Davidson ([email protected]) wrote:
: >
: >Again, you are a moron. I've spent years
: >negotiating license agreements with attorneys,
: >CEOs, and CFOs.
: >
: >> What the
: >> LGPL does is allow you to distribute such a derived work (and I quote)
: >> "under terms of your choice" [provided only that you allow reverse
: >> engineering and debugging and relinking and access to the source code
: >> of the Library, to summarize your obligations].
: >
: >Nope. Derivative work imposes restrictions and
: >demands on *your* code. Read the fucking agreement
: >again from end to end.

: While you were emptying the trash cans and ashtrays in these
: negotiating sessions, did anyone ever explain to you that if it
: is *not* a derivative work then *no* license is required at all?

: Clearly Peter is correct.

: >And never ever use any GNU-type license for
: >code that you want retain as your own.

: Unless of course you want to retain it, *and* be guaranteed free
: access to all derivative works.

The GPL does not guarantee free access to all derivative works.

The guarantees are only for the rights of the recipients to whom someone
distributes the software. If the original author is not a recipient of a
later modification then they cannot necessarily get the derivativations of
their code.

quote from preamble, emphasis mine

For example, if you distribute copies of such a program, whether
gratis or for a fee, you must give the __recipients__ all the
rights that you have. You must make sure that they, too, receive
or can get the source code.

In the version of COPYING that I am looking at the specific details are in
section 3. Clause a) allows you include the source code with the
distribution and in that case you explicitly do not have to make the
source code available in any other form to any other person, which means
that the only person guaranteed to get the source is the person to whom
the modified code is distributed.

I would suggest that other sections also imply this. In addition to the
preamble I already quoted, for example, section 6 quite explicitly
discusses the "recipient" receiving certain rights (i.e. the person to
whom the code is distributed), not simply any old person who might _like_
to be a recipient (but isn't for any reason).

The license does not say, as far as I can see, that distributing the code
to one person obligates you to distribute it to other people, except
potentially in section 3, part b, and c, but as I said, if you comply with
part a of section 3 (providing the recipient with the source as part of
the distribution) then you explicitly do not have to comply with part b or
c.

So in conclusion, if someone modifies your gpl'd code and sells it _with_
the source, then you cannot force them to give or sell the modified code
back to you. The sellers only obligation is to the purchaser to whom they
sold the software. (And that purchaser has no obligations to give
anything to anybody unless they choose to start distributing the code
themselves.)

 

[Floyd L. Davidson]

Floyd L. Davidson ([email protected]) wrote:
: >
: >Again, you are a moron. I've spent years
: >negotiating license agreements with attorneys,
: >CEOs, and CFOs.
: >
: >> What the
: >> LGPL does is allow you to distribute such a derived work (and I quote)
: >> "under terms of your choice" [provided only that you allow reverse
: >> engineering and debugging and relinking and access to the source code
: >> of the Library, to summarize your obligations].
: >
: >Nope. Derivative work imposes restrictions and
: >demands on *your* code. Read the fucking agreement
: >again from end to end.

: While you were emptying the trash cans and ashtrays in these
: negotiating sessions, did anyone ever explain to you that if it
: is *not* a derivative work then *no* license is required at all?

: Clearly Peter is correct.

: >And never ever use any GNU-type license for
: >code that you want retain as your own.

: Unless of course you want to retain it, *and* be guaranteed free
: access to all derivative works.

The GPL does not guarantee free access to all derivative works.

All derivative works in distribution then.

The guarantees are only for the rights of the recipients to whom someone
distributes the software. If the original author is not a recipient of a
later modification then they cannot necessarily get the derivativations of
their code.

quote from preamble, emphasis mine

For example, if you distribute copies of such a program, whether
gratis or for a fee, you must give the __recipients__ all the
rights that you have. You must make sure that they, too, receive
or can get the source code.

Which means if the derivative work is distributed, the original
author will have access, though perhaps indirectly.

....

So in conclusion, if someone modifies your gpl'd code and sells it _with_
the source, then you cannot force them to give or sell the modified code
back to you. The sellers only obligation is to the purchaser to whom they
sold the software. (And that purchaser has no obligations to give
anything to anybody unless they choose to start distributing the code
themselves.)

But you cannot *prevent* anyone who does have a copy from giving
(or selling) it to whoever they wish. The can't prevent is the
significant part.

 

[Todd Knarr]

Oh shit, i don't think Buffalo's product is open source. FSF should
bring them to the court.

IIRC the start of that was what happened. The offending parties decided
they didn't want to push it, hence why on Linksys' site you can download
the complete source code and build scripts for the firmware for their
products (except for the binary-only modules for the Broadcom network
chips, which were written in compliance with the Linux kernel's rules
for binary-only loadable kernel modules).

 

[justaguy]

The use of GPL'd code sure comes with a heavy price.

As opposed to the price of other copyrighted code?

 

[Shmuel (Seymour J.) Metz]

So my company is ok to include GPL/LGPL product in our product,
and we still can distribute our product in closed-source. But if i do
that, the court still have arguement?

The Devil is in the details. You need to involve your legal staff in
the packaging to ensure that you comply with the [L]GPL. I know of no
court case involving the GPL where there wasn't a real violation. I
see no reason to worry as long as you understand the rules and ensure
that they are followed.

--
Shmuel (Seymour J.) Metz, SysProg and JOAT <http://patriot.net/~shmuel>

Unsolicited bulk E-mail subject to legal action. I reserve the
right to publicly post or ridicule any abusive E-mail. Reply to
domain Patriot dot net user shmuel+news to contact me. Do not
reply to (e-mail address removed)

 

[Shmuel (Seymour J.) Metz]

OKOK , i understand now, the key point is "modified" and
"unmodified".

No! That's a secondary issue. If you modify it then you must provide
source code for the modification, but theat has *NOTHING* to do with
the dynamic versus static issue.

--
Shmuel (Seymour J.) Metz, SysProg and JOAT <http://patriot.net/~shmuel>

Unsolicited bulk E-mail subject to legal action. I reserve the
right to publicly post or ridicule any abusive E-mail. Reply to
domain Patriot dot net user shmuel+news to contact me. Do not
reply to (e-mail address removed)

 

[Shmuel (Seymour J.) Metz]

Why? Linking is one of the oldest and best defined concepts in
computer science. It's the pass in which symbolic references in the
code are replaced by addresses.

Or supplemented. There is no requirement to remove the names of the
symbolic references. IBM's Linkage Editor and Binder, for instance,
are fully capable[1] of relinking, including adding, deleting and
replacing sections.

[1] Unless someone specifies the not editable (NE) option, but
that's rare.

--
Shmuel (Seymour J.) Metz, SysProg and JOAT <http://patriot.net/~shmuel>

Unsolicited bulk E-mail subject to legal action. I reserve the
right to publicly post or ridicule any abusive E-mail. Reply to
domain Patriot dot net user shmuel+news to contact me. Do not
reply to (e-mail address removed)

 

[Shmuel (Seymour J.) Metz]

Linus opinion on this is irrelevant. Neither I nor the FSF nor
many others have released code under anything but the vanilla GPL.
By merging such code Linus lost his ability to vary the license.

What do you mean by "such code"? What code is Linus distributing under
the kernel GPL that he does not have the right to so distribute? Don't
confuse Linux with a Linux distribution.

--
Shmuel (Seymour J.) Metz, SysProg and JOAT <http://patriot.net/~shmuel>

Unsolicited bulk E-mail subject to legal action. I reserve the
right to publicly post or ridicule any abusive E-mail. Reply to
domain Patriot dot net user shmuel+news to contact me. Do not
reply to (e-mail address removed)

 

[Tor Iver Wilhelmsen]

What do you mean by "such code"? What code is Linus distributing under
the kernel GPL that he does not have the right to so distribute? Don't
confuse Linux with a Linux distribution.

Not to mention that the copyright holder is free to publish his works
under any (multiple) licenses they feel like, e.g. releasing a GPL
version of some program, and then release the same program (perhaps
with more features) under a commercial license, and for instance keep
working on the commercial codebase.

Case in point: Sendmail.

 

[Steve Schefter]

What do you mean by "such code"?

First of all, this is a quote from Alan Cox, so you'd have to ask him
what he meant by that to be sure. However, given the context (see my
orginal quote for the complete post), it appears to me that he's referring
to the bits of the code which he and others wrote and is now part of the
Linux kernel.

What code is Linus distributing under
the kernel GPL that he does not have the right to so distribute?

Neither Alan's quote (to my reading of it) nor my post make such a claim.
I was addressing what restrictions there might be on the licensing of
code writen for kernel add-ons. That is, if I write a kernel add-on,
might I have to put it under GPL too since the kernel is.

And before I get jumped on by folks reading this posting for saying that
I claim that all kernel add-ons must be under GPL, I offer no opinion. I
was simply replying to the note which indicated that "GPL or LGPL is NOT
viral" by saying that, where GPL in concerned at least, "there are differing
opinions".

Don't
confuse Linux with a Linux distribution.

While I realize that other postings in this thread talked of distributions,
I do not. I spoke strictly of the kernel and the kernel GPL.

Steve

 

[John Thompson]

["Followup-To:" header set to comp.os.linux.misc.]

Not to mention that the copyright holder is free to publish his works
under any (multiple) licenses they feel like, e.g. releasing a GPL
version of some program, and then release the same program (perhaps
with more features) under a commercial license, and for instance keep
working on the commercial codebase.

Case in point: Sendmail.

But sendmail is not licensed under the GPL:

http://www.sendmail.org/license-info.html

 

[Bill Unruh]

]In <[email protected]>, on 10/15/2004
] at 09:06 AM, (e-mail address removed) (Steve Schefter) said:

]> Linus opinion on this is irrelevant. Neither I nor the FSF nor
]> many others have released code under anything but the vanilla GPL.
]> By merging such code Linus lost his ability to vary the license.

]What do you mean by "such code"? What code is Linus distributing under
]the kernel GPL that he does not have the right to so distribute? Don't
]confuse Linux with a Linux distribution.

The code written by all of the people who contribute to the kernel.
Very little is actually written by Linus.
He has the right to distribute it because it is licensed to him under the
GPL. That means he has to abide by the conditions of the GPL just as anyone
else must-- unless each of those writers licensed it to him under some
other license.

 

[Tor Iver Wilhelmsen]

But sendmail is not licensed under the GPL:

http://www.sendmail.org/license-info.html

That wasn't the point, the point is that there are multiple licensing
options for it, at the authors' discretion.

 

[Malcolm Dew-Jones]

Bill Unruh ([email protected]) wrote:

: ]In <[email protected]>, on 10/15/2004
: ] at 09:06 AM, (e-mail address removed) (Steve Schefter) said:

: ]> Linus opinion on this is irrelevant. Neither I nor the FSF nor
: ]> many others have released code under anything but the vanilla GPL.
: ]> By merging such code Linus lost his ability to vary the license.

: ]What do you mean by "such code"? What code is Linus distributing under
: ]the kernel GPL that he does not have the right to so distribute? Don't
: ]confuse Linux with a Linux distribution.

: The code written by all of the people who contribute to the kernel.
: Very little is actually written by Linus.
: He has the right to distribute it because it is licensed to him under the
: GPL. That means he has to abide by the conditions of the GPL just as anyone
: else must

May yes, maybe no.

You might think that when he distributes code in the kernel that he did
not himself write, then yes, he would need to abide by the gpl, whereas
for those things which he wrote himself, no, he could take them out of the
current kernel and use them anyway he wants.

However the gpl really controls how those _contributors_ could distribute
linus's work after they modify it, not how the original author can
distribute the work after incorporating code that has been sent to him in
the hopes he finds it useful.

Many people who contributed did not do anything that could be considered a
"distribution" when they provided it, instead they "contributed" the code
directly to him, effectively making a donation, not a distribution - and
so for those sections of code, linux may very well be able to change the
license more than the contributors may realize. More importantly, many
contributions are modifications to code that linus originally wrote, and
copyright law is pretty clear that the original author would normally own
those modifications no matter who made them. For examples outside of
source code, consider if you rearrange and expand on someone elses short
story then the original author can claim copyright infringement - i.e.
they own the copyright on the new work because it is based on their work,
even though they didn't write most of it (there are lawsuits all the time
over film scripts that fall into this kind of category), or in music, if
you make an arrangement of a song then the original author explicitly owns
the copyright in the arrangement (even though they didn't write it).
Code is not inherently different, Linus likely does own the copyright in
much of the code that has been contributed, and can therefore distribute
it any way he chooses.

There will of course be situations where this would not be true, such as
any standalone utility that is not based on his code, but simply ties into
his code (and many contributions certainly fall into that category).

The above ignores the question of whether linus's code itself incorporates
gpl code that he has taken from distributions, effectively requiring him
to gpl everything no matter what anyways.

 

[Bill Unruh]

(e-mail address removed) (Malcolm Dew-Jones) writes:

]Bill Unruh ([email protected]) wrote:

]: ]In <[email protected]>, on 10/15/2004
]: ] at 09:06 AM, (e-mail address removed) (Steve Schefter) said:

]: ]> Linus opinion on this is irrelevant. Neither I nor the FSF nor
]: ]> many others have released code under anything but the vanilla GPL.
]: ]> By merging such code Linus lost his ability to vary the license.

]: ]What do you mean by "such code"? What code is Linus distributing under
]: ]the kernel GPL that he does not have the right to so distribute? Don't
]: ]confuse Linux with a Linux distribution.

]: The code written by all of the people who contribute to the kernel.
]: Very little is actually written by Linus.
]: He has the right to distribute it because it is licensed to him under the
]: GPL. That means he has to abide by the conditions of the GPL just as anyone
]: else must

]May yes, maybe no.

]You might think that when he distributes code in the kernel that he did
]not himself write, then yes, he would need to abide by the gpl, whereas
]for those things which he wrote himself, no, he could take them out of the
]current kernel and use them anyway he wants.

]However the gpl really controls how those _contributors_ could distribute
]linus's work after they modify it, not how the original author can
]distribute the work after incorporating code that has been sent to him in
]the hopes he finds it useful.

]Many people who contributed did not do anything that could be considered a
]"distribution" when they provided it, instead they "contributed" the code

?? what has "distribution" got to do with anything. They allowed Linus and
others to copy it. That is as far as their right goes. They can control
copying.

]directly to him, effectively making a donation, not a distribution - and
]so for those sections of code, linux may very well be able to change the
]license more than the contributors may realize. More importantly, many

No. Noone makes donations. Read the code.

]contributions are modifications to code that linus originally wrote, and
]copyright law is pretty clear that the original author would normally own
]those modifications no matter who made them. For examples outside of

No. Utterly false. Such a work is a derived work and copyright resides in BOTH
the original author and the person who changes it. It cannot be copied
except under conditions that BOTH agree to.


]source code, consider if you rearrange and expand on someone elses short
]story then the original author can claim copyright infringement - i.e.
]they own the copyright on the new work because it is based on their work,
]even though they didn't write most of it (there are lawsuits all the time
]over film scripts that fall into this kind of category), or in music, if
]you make an arrangement of a song then the original author explicitly owns
]the copyright in the arrangement (even though they didn't write it).

No he does not. He owns the copyright to the original work. He also jointly
with the arranger owns the copyright in the arrangement. Both have
copright and you need permission from both to copy.


]Code is not inherently different, Linus likely does own the copyright in
]much of the code that has been contributed, and can therefore distribute
]it any way he chooses.


Uh, perhaps you should read the kernel source code sometime. Almost all has
an explicit statement that the code is released under the GPL. It is NOT
"donated to Linus" nor is the copyright transfered to him.
Let me take a random bit of code (this is the first file in the kernel tree
I opened)

/*
* linux/drivers/ide/ide-dma.c Version 4.10 June 9, 2000
*
* Copyright (c) 1999-2000 Andre Hedrick <[email protected]>
* May be copied or modified under the terms of the GNU General Public
* License
*/


Note that Andre Hedrick retains the copyright, and allows others to use it
under the GPL.

 

[Malcolm Dew-Jones]

Bill Unruh ([email protected]) wrote:
: (e-mail address removed) (Malcolm Dew-Jones) writes:

: ]Many people who contributed did not do anything that could be considered a
: ]"distribution" when they provided it, instead they "contributed" the code

: ?? what has "distribution" got to do with anything. They allowed Linus and
: others to copy it. That is as far as their right goes. They can control
: copying.

: ]directly to him, effectively making a donation, not a distribution - and
: ]so for those sections of code, linux may very well be able to change the
: ]license more than the contributors may realize. More importantly, many


: No. Noone makes donations. Read the code.


From the linux kernel mail list faq

15.How do I get my patch into the kernel?

[part of the answer]

send it to linux-kernel and Cc: Linus Torvalds <[email protected]>
and hope Linus will apply it.
^^^^^^^^^^^^^^^^^^^^^^^^^^^^


I.e. to get something into the kernel you send it to Linus. The owner of
the code can certainly do that without themselves requiring any license
(because they own it).

So what makes you think that Linus must somehow get additional permission
(from the gpl in the code) to use the code? Linus already has permission
to use the code because the owner chose to send the code to him in the
hopes he would use it in the kernel.

The GPL you see in the code is part of the authors contribution to the
kernel. It is the authors way of giving permission to any third parties
who eventually receive the code as part of the kernel (which is what is
distributed eventually by Linus if he so choses). But it is not the
permission of the gpl in that code segment that allowed Linus to copy it
into the kernel in the first place - the author owned the code and chose
to send it to Linus explicitly for this very purpose, and it is that act
of sending the code to Linus that gives Linus permission to use it.