Newsgroups: comp.os.linux.announce
From: Ian Jackson < ijackson@nyx.cs.du.edu>
Subject: Copyright and licensing - a plea to software authors
Message-ID: < ann-13210.779119813@cs.cornell.edu>
Date: Fri, 9 Sep 1994 14:10:28 GMT
Approved: linux-announce@tc.cornell.edu (Matt Welsh)
If you are the author or maintainer of a piece of software intended to
be used under Linux please think very hard about what copyright notice
you attach to your work.
Many programs, often written especially for Linux, are being
distributed with copyright notices which are unclear, or prevent
distribution on CD-ROM, or often both.
Some even don't come with copyright notices and permission statements
at all, which means that all the users who download and use the
program in question are in violation of the author's copyright !
I would therefore strongly urge all members of the Linux development
community to consider whether using the GNU General Public Licence,
the Library GPL (LGPL) or perhaps a BSD UCB style copyright will serve
their goals better than a notice they hack up themselves.
Linus himself has said in an interview with the Linux Journal that
releasing Linux under the GPL was the best design decision he had
made.
Remember that if you put an awkward or unclear copyright notice on
your work you will be restricting the number of people who gain access
to your software.
A few myths dispelled:
* Releasing your software under the GPL does _not_ donate it to the
Free Software Foundation or relinquish _your_ rights to modify or
distribute it. It merely gives other people rights to do certain
things, and ensures that no-one can deny others those rights. For
example, Larry Wall has released Perl under both the GPL and his own
"Artistic Licence".
* The GPL does _not_ imply an interface copyright. It does _not_
require other people who write independent programs which run on or
with your software to release their work under the GPL. The statement
at the top of the COPYING file in the Linux source tree is redundant,
other than for clarity's sake and to silence ignorant flamers.
* Code which does not carry a copyright notice is _not_ public domain,
and people may _not_ do what they want with it. In fact, in the
absence of a statement saying what they may do or an explicit notice
saying the code is in the public domain they may not do _anything_
with it !
If you wish to place your work in the public domain, so that anyone
can do anything with it - including making their own derivations and
placing their own copyright on them without crediting you - you must
say so explicitly, for example with "I hereby relinquish my copyright
and place this work in the public domain".
Finally, note that in this posting I am _not_ seeking to limit your
right to put whatever copyright licence on your work you feel is
right. I'm trying to make you as a developer more conscious of the
ramifications of your decisions, so that you can make your own
decision about whether you want to benefit the Linux community and if
so how you wish to do so.
There is a FAQ on copyright law (mainly US law) which is posted to a
number of groups regularly including news.answers. It can be found on
rtfm.mit.edu in /pub/usenet/news.answers/Copyright-FAQ and in mirrors
thereof.
--
Ian Jackson, at home. ijackson@nyx.cs.du.edu or iwj10@cus.cam.ac.uk
+44 1223 575512 Escoerea on IRC. http://www.cl.cam.ac.uk/users/iwj10/
2 Lexington Close, Cambridge, CB4 3LS, England. Urgent: iwj@cam-orl.co.uk
--
Send submissions for comp.os.linux.announce to: linux-announce@tc.cornell.edu
Be sure to include Keywords: and a short description of your software.
Newsgroups: comp.os.linux.announce
From: "Theodore Ts'o" < tytso@MIT.EDU>
Subject: Re: Copyright and licensing - a plea to software authors
Message-ID: < ann-13962.779134356@cs.cornell.edu>
Date: Fri, 9 Sep 1994 18:12:57 GMT
Approved: linux-announce@tc.cornell.edu (Matt Welsh)
[Further discussion on this topic is directed to gnu.misc.discuss
and/or comp.os.linux.misc. These two messages have been posted to
c.o.l.announce to make the Linux community as a whole aware of
some issues surrounding the GPL; please see discussions elsewhere
if you wish to fight it out. --mdw]
From: Ian Jackson < ijackson@nyx.cs.du.edu>
Date: Fri, 9 Sep 1994 14:10:28 GMT
* The GPL does _not_ imply an interface copyright. It does _not_
require other people who write independent programs which run on or
with your software to release their work under the GPL. The statement
at the top of the COPYING file in the Linux source tree is redundant,
other than for clarity's sake and to silence ignorant flamers.
Not quite true. The FSF/Stallman interpretation of the GPL essentially
amounts to an interface copyright, although they attempt to deny it. It
is very important that prospective authors understand this before using
the GPL. It isn't a problem, as long as you and not the FSF control the
copyright. But in some cases you *do* need to put some disclaimers in
the COPYING file such as is done in the Linux source tree.
It the FSF's claim that it you can not write your own program that uses
the gmp interface, and distribute it using a copyright notice different
than the GPL. Despite the fact your distribution doesn't contain a
character of GPL, but just merely links against the gmp library, the FSF
and Stallman will still claim that your distribution violates the GPL
copyright on the gmp library.
Stallman and FSF also threatened NeXT with similar GPL copyright claims
to make NeXT release the Objective C front-end under the GPL, despite
the fact that the front-end originally didn't contain a line of GPL'ed
code, but merely called the GPL'ed GNU gcc compiler. Arguably, the
"right thing" happened in that the Objective C compiler became freely
available. However, it should be noted that the means to this end was
in essence a threat based on interface copyright. (I have email from
Stallman to back up this story; I will forward it to people upon
request.)
Now, a number of law professors have stated that the FSF would be
"laughed out of court" if they ever tried to sue someone over this. But
it is still true that the FSF's interpretation has caused a lot of Fear,
Uncertainty, and Doubt.
This is the reason for the COPYING file in the top of the Linux source
tree, and the reason why Wine was not signed over to the FSF, despite
the FSF's request to the Wine developers. (After all, Windows uses
dynamic libraries extensively, and according to the FSF interpretation,
anything that dynamic links to GPL code must fall under the GPL; since
in order to run programs like Microsoft Excel and Word, they would have
to dynamic link to GPL'ed code in the Wine emulator, the FSF
interpretation would preclude the use of such commercially available
programs under Wine!)
This being said, please keep in mind that this problem with the GPL is
only a problem with the FSF's *interpretation* of the GPL. It is not a
problem with the GPL itself! As long as you, the author, retain control
over the Copyright, it is up to you to decide how you want to interpret
and enforce the Copyright. You can certainly decide to release things
under terms which are more liberal that the GPL. For example, Perl is
released under either the terms of the GPL, or Larry Wall's "Artistic
License", at the descretion of the user.
At the very least, if you are releasing a library under the GPL, please
consider whether you would like it to come under the GPL or the LGPL,
and if it is under the GPL, you may wish to disclaim any interface
copyrights, and explicitly allow poeple to write and distribute code
which calls your GPL'ed routines, as long as their code doesn't actually
include the GPL'ed routines in the distribution, but rather uses them by
reference.
If your routines are released as a Linux shared library, you should
consider allowing people to write and distribute source and binaries
which link against the shared library, even if their programs don't fall
under the GPL. Since the FSF believes that any program which
dynamically links against GPL'ed code must also fall under the GPL, if
you wish to allow people to release binaries which link against your
GPL'ed version of curses (for example) it would be wise to explicitly
state that it is O.K. for people to do this.
As the author, you have many rights, and it is definitely a Good Thing
to exercise them wisely.
- Ted
--
Send submissions for comp.os.linux.announce to: linux-announce@tc.cornell.edu
Be sure to include Keywords: and a short description of your software.
Path: bga.com!news.sprintlink.net!sundog.tiac.net!usenet.elf.com!
news2.near.net!MathWorks.Com!yeshua.marcam.com!usc!howland.reston.ans.net!
pipex!lyra.csx.cam.ac.uk!iwj10
From: iw...@cus.cam.ac.uk (Ian Jackson)
Newsgroups: comp.os.linux.misc,gnu.misc.discuss
Subject: Re: Copyright and licensing - a plea to software authors
Followup-To: comp.os.linux.misc
Date: Sat, 10 Sep 1994 00:59:47 GMT
Organization: Linux Unlimited
Lines: 48
Message-ID: <1994Sep10.005947.4890.chiark.ijackson@nyx.cs.du.edu>
References: <ann-13210.779119813@cs.cornell.edu>
<ann-13962.779134356@cs.cornell.edu>
NNTP-Posting-Host: bootes.cus.cam.ac.uk
Summary: Words fail me.
Keywords: GPL, copyright
Originator: iw...@bootes.cus.cam.ac.uk
In article <ann-13962.779134...@cs.cornell.edu>,
Theodore Ts'o <ty...@MIT.EDU> wrote:
>[Further discussion on this topic is directed to gnu.misc.discuss
>and/or comp.os.linux.misc. These two messages have been posted to
>c.o.l.announce to make the Linux community as a whole aware of
>some issues surrounding the GPL; please see discussions elsewhere
>if you wish to fight it out. --mdw]
I AM ABSOLUTELY LIVID. FURIOUS DOESN'T EVEN BEGIN TO DESCRIBE IT.
Firstly, because Ted Ts'o has joined in with the ignorant flamers. He
himself has stooped to spreading the Fear, Uncertainty and Doubt that
he claims he finds so harmful.
There was nothing inaccurate about my posting, and nothing that needed
correcting. The posting went through several drafts on a mailing list
with several hundred people, with _no_ objections.
Secondly, because the moderator of comp.os.linux.announce, Matt Welsh,
posted this so-called "correction" WITHOUT ANY REFERENCE TO ME. I
find this particularly galling considering the fact that Matt is a
*member* of the mailing list I mention above.
Unfortunately I am going away on holiday for a week, so I will not be
able to battle this out.
However, I am EXTREMELY ANGRY that Theodore Ts'o has turned a message
that would have accurately and honestly reassured those who have been
confused by the FUD surrounding this topic into an opportunity for him
to score points of the FSF and RMS, and that Matt Welsh, whom I would
have expected to know better, allowed him to do so.
This has greatly damaged the cause of free software, and will no doubt
make the lives of all the developers and users of Linux much more
difficult as they wrestle with the problems caused by missing,
unclear, home-built or otherwise unhelpful copyright notices.
>Now, a number of law professors have stated that the FSF would be
>"laughed out of court" if they ever tried to sue someone over this. But
>it is still true that the FSF's interpretation has caused a lot of Fear,
>Uncertainty, and Doubt.
Which YOU are helping to spread.
--
Ian Jackson, at home. ijack...@nyx.cs.du.edu or iw...@cus.cam.ac.uk
+44 1223 575512 Escoerea on IRC. http://www.cl.cam.ac.uk/users/iwj10/
2 Lexington Close, Cambridge, CB4 3LS, England. Urgent: i...@cam-orl.co.uk
Path: nntp.gmd.de!xlink.net!howland.reston.ans.net!europa.eng.gtefsd.com!
news.umbc.edu!haven.umd.edu!nova.umd.edu!nova!rockwell
From: rock...@nova.umd.edu (Raul Deluth Miller)
Newsgroups: comp.os.linux.misc
Subject: Re: Copyright and licensing - a plea to software authors
Followup-To: gnu.misc.discuss
Date: 11 Sep 1994 18:07:00 -0400
Organization: University of Maryland University College
Lines: 91
Distribution: world
Message-ID: <ROCKWELL.94Sep11180659@nova.umd.edu>
References: <ann-13210.779119813@cs.cornell.edu>
<ann-13962.779134356@cs.cornell.edu>
NNTP-Posting-Host: nova.umd.edu
In-reply-to: "Theodore Ts'o"'s message of Fri, 9 Sep 1994 18:12:57 GMT
Theodore Ts'o:
: The FSF/Stallman interpretation of the GPL essentially amounts to
: an interface copyright, although they attempt to deny it.
:
: ... (I have email from Stallman to back up this story; I will
: forward it to people upon request.)
Attached is the letter that Ts'o forwarded to me. I think that it
paints a significantly different picture from what Ts'o described.
But read it for yourself.
--
Raul D. Miller n =: p*q NB. 9<##:##:n [.large prime p, q
<rock...@nova.umd.edu> y =: n&|&(*&x)^:e 1 NB. -.1 e.e e.&factors<:p,q [.e<n
NB. public e, n, y
x -: n&|&(*&y)^:d 1 NB. 1=(d*e)+.p*&<:q
attachment: forwarded mail
Date: Wed, 7 Jul 93 01:25:07 -0400
From: r...@gnu.ai.mit.edu (Richard Stallman)
To: ty...@Athena.MIT.EDU
In-Reply-To: <930707042...@tsx-11.MIT.EDU> (ty...@ATHENA.MIT.EDU)
Subject: Re: Fascist GPL interpretation?
If it is not a new position, could you please supply some references?
That would be greatly appreciated.
Sorry, I don't understand.
Again, I'd like to know what your reasons are.
Here's the explanation I wrote (though I would have expected you'd
have seen it by now).
Several years ago, I met with Steve Jobs, who was looking for some
alternative to making the Objective C front end free software. (This
may have been due to worries about being hassled by Stepstone, rather
than a desire to be uncooperative.) He asked me if it it would be
legal to ship proprietary .o files to the user and have the user link
them with the GNU compiler.
At that time, I envisaged the legal ramifications like some others who
have recently posted on this list, so I did not see a basis for saying
they could not do this.
But at the same time, I realized that it would not bode well for the
GNU project if such a thing were permitted. So I responded, "I will
have to check with our lawyer."
It's a good thing I did, because when I checked, I found that there
was a basis for objecting to this plan. Such .o files would have
implied the presence of the GNU compiler, linked with them. They
would be, in effect, a way of distributing a larger program which
implicitly includes the GNU compiler; as such, it must follow the
terms on the GNU compiler.
I told NeXT this, and NeXT decided there was no alternative to making
the Objective C front end free software. So now it is available to
all of us as a part of GCC.
Note that this is not a matter of copyrighting an interface. The .o
files that NeXT planned to release would have used one of the
(internal) interfaces of the GNU compiler, but that was *not* what the
FSF objected to. Our objection was because the use of these .o files
implied linking them with the GNU compiler--the program, not just an
interface.
If it were possible for a company to get around the GPL simply by
dressing up changes or extensions as "separate programs that the user
might link in", then the GPL would be a paper tiger. (True, we often
print it on paper, but...) So it is vital for the FSF to object.
If we made this a request rather than a legal demand, some people
would comply as a matter of conscience. But many others who would
not. If I had told Jobs, "It is legal, but please don't," I doubt he
would have heeded the request.
Many improvements to GNU software are contributed by, or funded by,
companies. If they could make these improvements proprietary, many of
them would. I consider these companies unethical (because making
proprietary software is unethical in general), but they don't share my
ethical views, and they don't feel they should forego profit for a
mere request from the FSF.
The only way to make sure these improvements are free software is to
make it hard to make them proprietary. That's what the GPL is for,
and to make it work right, we must not permit getting around it by
"having the user do the link."
Newsgroups: comp.os.linux.misc
Path: bga.com!news.sprintlink.net!uunet!usc!howland.reston.ans.net!
europa.eng.gtefsd.com!MathWorks.Com!news.kei.com!travelers.mail.cornell.edu!
cornell!mdw
From: m...@cs.cornell.edu (Matt Welsh)
Subject: Re: Copyright and licensing - a plea to software authors
Message-ID: <1994Sep14.141220.28892@cs.cornell.edu>
Keywords: GPL, copyright
Organization: Cornell CS Robotics and Vision Laboratory, Ithaca, NY 14850
References: <ann-13210.779119813@cs.cornell.edu>
<ann-13962.779134356@cs.cornell.edu>
<1994Sep10.005947.4890.chiark.ijackson@nyx.cs.du.edu>
Date: Wed, 14 Sep 1994 14:12:20 GMT
Lines: 30
In article <1994Sep10.005947.4890.chiark.ijack...@nyx.cs.du.edu>
iw...@cus.cam.ac.uk (Ian Jackson) writes:
>I AM ABSOLUTELY LIVID. FURIOUS DOESN'T EVEN BEGIN TO DESCRIBE IT.
All right, Ian. Calm down. It is, after all, only USENET.
>Secondly, because the moderator of comp.os.linux.announce, Matt Welsh,
>posted this so-called "correction" WITHOUT ANY REFERENCE TO ME. I
>find this particularly galling considering the fact that Matt is a
>*member* of the mailing list I mention above.
I've already mailed you about this, but here's my defense: I
decided to approve Ted's posting to present an "alternate viewpoint"
to the opinions (yes, opinions) stated in your article. In retrospect,
neither article should have been posted to c.o.l.announce. But I'm
having to put on the moderator hat, so I can't play favourites.
Get me? There were a number of impassioned responses both to your
article as well as Ted's, but I refused to post them, because c.o.l.a
shouldn't be a mudslinging match between Linux enthusiasts.
Both articles were accurate as to their claims. The rest is opinion.
I'm not going to defend my decisions any further. Maybe I screwed up. Cope.
>This has greatly damaged the cause of free software, and will no doubt
>make the lives of all the developers and users of Linux much more
>difficult as they wrestle with the problems caused by missing,
>unclear, home-built or otherwise unhelpful copyright notices.
You have a gift for overstatement, Ian.
M. Welsh
Path: bga.com!news.sprintlink.net!sundog.tiac.net!usenet.elf.com!
news2.near.net!bloom-beacon.mit.edu!senator-bedfellow.mit.edu!
athena.mit.edu!tytso
From: ty...@athena.mit.edu (Theodore Y. Ts'o)
Newsgroups: comp.os.linux.misc
Subject: Re: Copyright and licensing - a plea to software authors
Followup-To: comp.os.linux.misc
Date: 14 Sep 1994 15:41:10 GMT
Organization: Massachusetts Institute of Technology
Lines: 53
Message-ID: <TYTSO.94Sep14114127@dcl.mit.edu>
References: <ann-13210.779119813@cs.cornell.edu> <ann-13962.779134356@cs.cornell.edu>
<1994Sep10.005947.4890.chiark.ijackson@nyx.cs.du.edu>
NNTP-Posting-Host: dcl.mit.edu
In-reply-to: iwj10@cus.cam.ac.uk's message of Sat, 10 Sep 1994 00:59:47 GMT
In article <1994Sep10.005947.4890.chiark.ijack...@nyx.cs.du.edu>
iw...@cus.cam.ac.uk (Ian Jackson) writes:
>I AM ABSOLUTELY LIVID. FURIOUS DOESN'T EVEN BEGIN TO DESCRIBE IT.
>Firstly, because Ted Ts'o has joined in with the ignorant flamers. He
>himself has stooped to spreading the Fear, Uncertainty and Doubt that
>he claims he finds so harmful.
I'm sorry you feel that way. But your posting was simply inaccurate
regarding the issue of interface copyright. It is a fact that Stallman
stomped on a source distribution of PGP that merely contained #ifdef's
so that it could be potentially linked with the gmp library, which was
protected by the GPL. Sure, Stallman and the FSF couched it in very
legalistic terms, and then justified the means with the ends ("Free
Software is Good, so what we did is O.K."). But that's not any
different from what Lotus did when they tried to use Copyright to
protect their user interface. It was the same sort of legal mumbo-jumbo
that led to the same sort of result --- a restriction of who can write
independent programs that implement or use a particular interface,
whether that be a UI or an API.
I am very saddened by the FSF's decision to use this very broad
interpretation of the GPL. I engaged in several e-mail messages
debating this issue with Stallman, over a year ago, when the PGP/gmp
issue came up, and he would not be moved. He believes he is right, and
he is entitled to that belief.
However, my sadness does not extend to trying to ignore the situation
for what it is. I will not sweep this situation under the rug. People
are entitled to know *all* of the issues of the GPL, both its good side
and its shadow side.
However, I would like to reassure people that this has not prevented me
from writing and releasing things that fall under the GPL. I've written
several programs which I've released under the GPL, and I encourage
others to do so. However, I would have to think twice before deciding
to release the ownership of any significant amount of code to the FSF,
because of their interpretation of the GPL. As long as I own the
copyright, I get to control how the GPL is interpreted, and I have
lattitude over granting exceptions to the GPL as I see fit. If I sign
my code over to the FSF, I lose this control, and right now, I don't
trust the FSF enough to do that. They have too much of a political
agenda which they are trying to push.
Every author has the right to make this decision for him or herself. I
don't say that my views are the only right one. However, I believe that
everyone is entitled to all of the facts, and you, Ian, did not include
all of those facts in an attempt to "reassure" people.
I believe the cause of free software is best furthered by putting the
truth --- the whole truth --- on the table, and letting the chips fall
where they may.
- Ted
Path: bga.com!news.sprintlink.net!howland.reston.ans.net!gatech!ncar!
csn!ub!ns.potsdam.edu!news.potsdam.edu!nelson
From: nel...@crynwr.crynwr.com (Russell Nelson)
Newsgroups: comp.os.linux.misc,gnu.misc.discuss
Subject: Re: Copyright and licensing - a plea to software authors
Followup-To: gnu.misc.discuss
Date: 16 Sep 1994 15:41:31 GMT
Organization: Crynwr Software
Lines: 37
Message-ID: <NELSON.94Sep16114131@crynwr.crynwr.com>
References: <ann-13210.779119813@cs.cornell.edu> <ann-13962.779134356@cs.cornell.edu>
<1994Sep10.005947.4890.chiark.ijackson@nyx.cs.du.edu>
<TYTSO.94Sep14114127@dcl.mit.edu>
NNTP-Posting-Host: nh2.potsdam.edu
In-reply-to: tytso@athena.mit.edu's message of 14 Sep 1994 15:41:10 GMT
In article <TYTSO.94Sep14114...@dcl.mit.edu> ty...@athena.mit.edu
(Theodore Y. Ts'o) writes:
I'm sorry you feel that way. But your posting was simply inaccurate
regarding the issue of interface copyright. It is a fact that Stallman
stomped on a source distribution of PGP that merely contained #ifdef's
so that it could be potentially linked with the gmp library, which was
protected by the GPL. Sure, Stallman and the FSF couched it in very
legalistic terms, and then justified the means with the ends ("Free
Software is Good, so what we did is O.K."). But that's not any
different from what Lotus did when they tried to use Copyright to
protect their user interface.
I don't think you understand Ted. At least, that is all I can assume
from your paragraph above. There *is* a difference. If someone
invents an interface between two packages, for example taking a piece
of GPL'ed code and making it into a subroutine package, then letting
the user link that code into a program, THAT violates the GPL.
The whole issue, to me, depends on whether or not the interface is
created solely to work around the GPL, or if it was created for other
purposes. If this distinction is not made, if, say, the GPL must
apply to any code that dynamically links into GPL'ed code, then
GPL'ing code becomes much less attractive. For example, my GPL'ed packet
drivers dynamically link into the MS-DOS kernel. Obviously the GPL
cannot be made to apply to MS-DOS, so I would not be able to use the
GPL on my code.
If, on the other hand, anyone may create a dynamic link to a GPL
package, voiding the GPL, even if the dynamic link was ONLY created to
avoid the GPL, then the GPL has little force, and one may as well put
code into the public domain.
--
-russ <nel...@crynwr.com> http://www.crynwr.com/crynwr/nelson.html
Crynwr Software | Crynwr Software sells packet driver support | ask4 PGP key
11 Grant St. | +1 315 268 1925 (9201 FAX) | What is thee doing about it?
Potsdam, NY 13676 | LPF member - ask me about the harm software patents do.
Path: bga.com!news.sprintlink.net!howland.reston.ans.net!europa.eng.gtefsd.com!
MathWorks.Com!news2.near.net!bloom-beacon.mit.edu!senator-bedfellow.mit.edu!
senator-bedfellow.mit.edu!not-for-mail
From: "Theodore Ts'o" <ty...@MIT.EDU>
Newsgroups: comp.os.linux.misc
Subject: Re: Copyright and licensing - a plea to software authors
Date: 24 Sep 1994 22:53:02 -0400
Organization: The Internet
Lines: 62
Sender: n...@athena.mit.edu
Distribution: world
Message-ID: <362ome$6cp@senator-bedfellow.MIT.EDU>
Reply-To: ty...@MIT.EDU
NNTP-Posting-Host: senator-bedfellow.mit.edu
From: nel...@crynwr.crynwr.com (Russell Nelson)
Crossposted-To: gnu.misc.discuss
Date: 16 Sep 1994 15:41:31 GMT
I don't think you understand Ted. At least, that is all I can assume
from your paragraph above. There *is* a difference. If someone
invents an interface between two packages, for example taking a piece
of GPL'ed code and making it into a subroutine package, then letting
the user link that code into a program, THAT violates the GPL.
The whole issue, to me, depends on whether or not the interface is
created solely to work around the GPL, or if it was created for other
purposes. If this distinction is not made, if, say, the GPL must
apply to any code that dynamically links into GPL'ed code, then
GPL'ing code becomes much less attractive. For example, my GPL'ed packet
drivers dynamically link into the MS-DOS kernel. Obviously the GPL
cannot be made to apply to MS-DOS, so I would not be able to use the
GPL on my code.
If, on the other hand, anyone may create a dynamic link to a GPL
package, voiding the GPL, even if the dynamic link was ONLY created to
avoid the GPL, then the GPL has little force, and one may as well put
code into the public domain.
This is precisely the hypocrisy that I'm complaining about. It seems to
me that you (and the FSF) want to have it both ways. Well, I'm sorry,
but if PGP and gmp are considered "one program", then your drivers and
MS-DOS must also be considered "one program". You can't have it both
ways.
It seems to me that people are making a distinction as a convenient way
to control the outcome of how they want things to come out. If they
want it to be allowed, then "obviously" the GPL cannot be made to apply
to MS-DOS. But if they don't, then "obviously" the GPL must apply.
Sorry, but the legal system doesn't work that way.
Another example --- suppose I write a program that uses dbm; it can
potentially be linked against gdbm. Hence, by your reasoning, my
program must fall under the GPL! But perhaps the fact that there is a
non-GPL library is enough to make it O.K. Alright, I'll write a slow,
stub library which implements the gmp interface. Then PGP must be OK!
A stub library isn't enough? Alright, I'll write a library which
implements the gmp interface but calls a slower package as its back-end.
Now is that OK? I'm sure the FSF would find some reason why that
wouldn't be OK, since they dislike PGP so much.
The point at which something becomes OK by the FSF's "definition" is
purely arbitrary, which is what I dislike. There is an entirely
separate question which is whether or not the FSF interpretation would
possibly even hold water in a court of law, or whether the FSF would be
laughed out of court. Short of a test case actually coming before a
court, we won't know for certain the answer to this.
But even if the FSF interpretation is legally airtight, the fact that it
is arbitrary and depends on what is situationally convenient disturbs
me. Fortunately, as long as you and I, the authors, own the copyright
on the code, and not the FSF, this trumps the entire issue. This is why
I suggest that authors think twice before donating the ownership of
their code to the FSF.
- Ted
Path: bga.com!news.sprintlink.net!howland.reston.ans.net!europa.eng.gtefsd.com!
MathWorks.Com!news2.near.net!bloom-beacon.mit.edu!senator-bedfellow.mit.edu!
senator-bedfellow.mit.edu!not-for-mail
From: nel...@crynwr.com (Russell Nelson)
Newsgroups: comp.os.linux.misc
Subject: Re: Copyright and licensing - a plea to software authors
Date: 28 Sep 1994 23:10:20 -0400
Organization: The Internet
Lines: 45
Sender: n...@athena.mit.edu
Distribution: world
Message-ID: <36db6s$8bp@senator-bedfellow.MIT.EDU>
Reply-To: nel...@crynwr.com (Russell Nelson)
NNTP-Posting-Host: senator-bedfellow.mit.edu
Date: Sat, 24 Sep 94 22:54:08 -0400
From: Theodore Ts'o <ty...@MIT.EDU>
Cc: Linux-M...@senator-bedfellow.MIT.EDU
Another example --- suppose I write a program that uses dbm; it can
potentially be linked against gdbm. Hence, by your reasoning, my
program must fall under the GPL!
No, not at all, never, no way. Your program uses dbm. dbm has a
known interface. Just because you *can* use gdbm, that doesn't put
your program under the GPL.
But perhaps the fact that there is a non-GPL library is enough to
make it O.K. Alright, I'll write a slow, stub library which
implements the gmp interface. Then PGP must be OK! A stub library
isn't enough? Alright, I'll write a library which implements the
gmp interface but calls a slower package as its back-end. Now is
that OK? I'm sure the FSF would find some reason why that wouldn't
be OK, since they dislike PGP so much.
It's not as mechanistic as that. If you wrote the stub library as a
way to let the user do the link to a GPL'ed package, then you aren't
accomplishing anything.
The point at which something becomes OK by the FSF's "definition" is
purely arbitrary, which is what I dislike.
Then what you wish to escape is the legal system. Fine, don't use a
copyright.
There is an entirely separate question which is whether or not the
FSF interpretation would possibly even hold water in a court of
law, or whether the FSF would be laughed out of court. Short of a
test case actually coming before a court, we won't know for certain
the answer to this.
Judges aren't stupid, and they don't like it when you try to fool
them. If an attorney could show the court that any of the above
coding is a subterfuge intended to misappropriate a GPL'ed program,
the court will find in his favor.
-russ <nel...@crynwr.com> http://www.crynwr.com/crynwr/nelson.html
Crynwr Software | Crynwr Software sells packet driver support | ask4 PGP key
11 Grant St. | +1 315 268 1925 (9201 FAX) | What is thee doing about it?
Potsdam, NY 13676 | LPF member - ask me about the harm software patents do.
Path: bga.com!news.sprintlink.net!howland.reston.ans.net!europa.eng.gtefsd.com!
MathWorks.Com!news2.near.net!bloom-beacon.mit.edu!senator-bedfellow.mit.edu!
senator-bedfellow.mit.edu!not-for-mail
From: "Theodore Ts'o" <ty...@MIT.EDU>
Newsgroups: comp.os.linux.misc
Subject: Re: Copyright and licensing - a plea to software authors
Date: 28 Sep 1994 23:36:08 -0400
Organization: The Internet
Lines: 49
Sender: n...@athena.mit.edu
Distribution: world
Message-ID: <36dcn8$92i@senator-bedfellow.MIT.EDU>
Reply-To: ty...@MIT.EDU
NNTP-Posting-Host: senator-bedfellow.mit.edu
Followup-to: gnu.misc.discuss
Date: Wed, 28 Sep 94 23:14 EDT
From: nel...@crynwr.com (Russell Nelson)
There is an entirely separate question which is whether or not the
FSF interpretation would possibly even hold water in a court of
law, or whether the FSF would be laughed out of court. Short of a
test case actually coming before a court, we won't know for certain
the answer to this.
Judges aren't stupid, and they don't like it when you try to fool
them. If an attorney could show the court that any of the above
coding is a subterfuge intended to misappropriate a GPL'ed program,
the court will find in his favor.
As I said before, short of a test case actually coming before a court,
we won't know that. The law is the law, and if copyright law doesn't
happen to be convenient for the Free Software Foundation, that's just
too bad. The fact that certain laws don't work they way you would like
isn't a case of "subterfuge". So I don't find your rationale (which I
think is similar or the same to Stallman's rationale --- did you just
use his words?) pursuasive.
In any case, that's not the important issue. By merely trying to
prohibit someone from distribute a program that's coded to a particular
interface, even though every single line of code in that program is
written BY THAT PERSON, then the you and the FSF are in effect trying to
assert what might as well be an interface copyright. In effect, there
is an attempt using copyright law to try to put restrictions on software
coded to a particular interface --- at least in the case of any program
written to use the gmp interface.
Sure, you have a great justification for it, which is that it helps
promote the FSF's agenda of its particular vision of free software, but
that's a means justify the ends argument. Even if you think it is a
justified form of interface copyright, it's still a form of interface
copyright.
Whether or not the FSF's attempt at this interface copyright would hold
water is a different question, and it's not worth argueing here, since
neither of us will know until it comes before a judge and jury --- and
it's probably in the best interests of the FSF for it not to actually
come into a courtroom setting anyway.
But the mere fact that the FSF is even trying to do this is something
that I find morally repugnant.
- Ted
Path: bga.com!news.sprintlink.net!howland.reston.ans.net!spool.mu.edu!
bloom-beacon.mit.edu!senator-bedfellow.mit.edu!senator-bedfellow.mit.edu!
not-for-mail
From: nel...@crynwr.com (Russell Nelson)
Newsgroups: comp.os.linux.misc
Subject: Re: Copyright and licensing - a plea to software authors
Date: 28 Sep 1994 23:48:53 -0400
Organization: The Internet
Lines: 26
Sender: n...@athena.mit.edu
Distribution: world
Message-ID: <36ddf5$99t@senator-bedfellow.MIT.EDU>
Reply-To: nel...@crynwr.com (Russell Nelson)
NNTP-Posting-Host: senator-bedfellow.mit.edu
Date: Wed, 28 Sep 1994 23:36:14 +0500
From: Theodore Ts'o <ty...@MIT.EDU>
In any case, that's not the important issue. By merely trying to
prohibit someone from distribute a program that's coded to a particular
interface, even though every single line of code in that program is
written BY THAT PERSON, then the you and the FSF are in effect trying to
assert what might as well be an interface copyright. In effect, there
is an attempt using copyright law to try to put restrictions on software
coded to a particular interface --- at least in the case of any program
written to use the gmp interface.
An odd kind of interface copyright if you can unilaterally take the
"copyright" away from the "copyright holder", by actually programming
to that interface.
In other words, if you don't like the gmp "interface copyright", write
a package that is compatible with it (that people would seriously
use). That act takes away the "interface copyright", which could not
happen under copyright law if an actual copyright on the interface was
claimed.
-russ <nel...@crynwr.com> http://www.crynwr.com/crynwr/nelson.html
Crynwr Software | Crynwr Software sells packet driver support | ask4 PGP key
11 Grant St. | +1 315 268 1925 (9201 FAX) | What is thee doing about it?
Potsdam, NY 13676 | LPF member - ask me about the harm software patents do.
Path: bga.com!news.sprintlink.net!howland.reston.ans.net!spool.mu.edu!
bloom-beacon.mit.edu!senator-bedfellow.mit.edu!senator-bedfellow.mit.edu!
not-for-mail
From: "Theodore Ts'o" <ty...@MIT.EDU>
Newsgroups: comp.os.linux.misc
Subject: Re: Copyright and licensing - a plea to software authors
Date: 28 Sep 1994 23:58:50 -0400
Organization: The Internet
Lines: 27
Sender: n...@athena.mit.edu
Distribution: world
Message-ID: <36de1q$9h2@senator-bedfellow.MIT.EDU>
Reply-To: ty...@MIT.EDU
NNTP-Posting-Host: senator-bedfellow.mit.edu
Date: Wed, 28 Sep 94 23:53 EDT
From: nel...@crynwr.com (Russell Nelson)
An odd kind of interface copyright if you can unilaterally take the
"copyright" away from the "copyright holder", by actually programming
to that interface.
In other words, if you don't like the gmp "interface copyright", write
a package that is compatible with it (that people would seriously
use). That act takes away the "interface copyright", which could not
happen under copyright law if an actual copyright on the interface was
claimed.
So whether or not package A can be distrbuted only under the terms
attached to package B depends on the existence or non-existence of
package C, where A, B, and C do not share any lines of codes and are not
otherwise derived from one another?
This is rational?
I suppose that since no one else has written a freeware distribution of
MS-DOS, the fact that your drivers dynamically link with MS-DOS means
that they are "one program", and you are therefore misappropriating
Microsloth's program by using the subterfuge of distributing drivers
separately from MS-DOS?
- Ted
Path: bga.com!news.sprintlink.net!howland.reston.ans.net!gatech!
newsxfer.itd.umich.edu!zip.eecs.umich.edu!yeshua.marcam.com!MathWorks.Com!
news2.near.net!bloom-beacon.mit.edu!senator-bedfellow.mit.edu!
senator-bedfellow.mit.edu!not-for-mail
From: "Theodore Ts'o" <ty...@MIT.EDU>
Newsgroups: comp.os.linux.misc
Subject: Re: Copyright and licensing - a plea to software authors
Date: 29 Sep 1994 13:32:11 -0400
Organization: The Internet
Lines: 18
Sender: n...@athena.mit.edu
Distribution: world
Message-ID: <36etmr$114@senator-bedfellow.MIT.EDU>
Reply-To: ty...@MIT.EDU
NNTP-Posting-Host: senator-bedfellow.mit.edu
Date: Thu, 29 Sep 94 10:05 EDT
From: nel...@crynwr.com (Russell Nelson)
So whether or not package A can be distrbuted only under the
terms attached to package B depends on the existence or
non-existence of package C, where A, B, and C do not share any
lines of codes and are not otherwise derived from one another?
This is rational?
Sure, because the copyright for B says that if you use it, you're
making a derivation, and copyright law says that you can control the
use of derivations.
Sorry, the definition of "dervitive work" is defined in the copyright law,
not in the GPL. So is the definition of "fair use".
- Ted
Path: bga.com!news.sprintlink.net!howland.reston.ans.net!spool.mu.edu!
bloom-beacon.mit.edu!senator-bedfellow.mit.edu!senator-bedfellow.mit.edu!
not-for-mail
From: nel...@crynwr.com (Russell Nelson)
Newsgroups: comp.os.linux.misc
Subject: Re: Copyright and licensing - a plea to software authors
Date: 29 Sep 1994 10:02:19 -0400
Organization: The Internet
Lines: 46
Sender: n...@athena.mit.edu
Distribution: world
Message-ID: <36ehdb$oe2@senator-bedfellow.MIT.EDU>
Reply-To: nel...@crynwr.com (Russell Nelson)
NNTP-Posting-Host: senator-bedfellow.mit.edu
Date: Wed, 28 Sep 1994 23:59:02 +0500
From: Theodore Ts'o <ty...@MIT.EDU>
Cc: Linux-M...@senator-bedfellow.MIT.EDU
Address: 1 Amherst St., Cambridge, MA 02139
Phone: (617) 253-8091
Date: Wed, 28 Sep 94 23:53 EDT
From: nel...@crynwr.com (Russell Nelson)
An odd kind of interface copyright if you can unilaterally take the
"copyright" away from the "copyright holder", by actually programming
to that interface.
In other words, if you don't like the gmp "interface copyright", write
a package that is compatible with it (that people would seriously
use). That act takes away the "interface copyright", which could not
happen under copyright law if an actual copyright on the interface was
claimed.
So whether or not package A can be distrbuted only under the terms
attached to package B depends on the existence or non-existence of
package C, where A, B, and C do not share any lines of codes and are not
otherwise derived from one another?
This is rational?
Sure, because the copyright for B says that if you use it, you're
making a derivation, and copyright law says that you can control the
use of derivations.
What you're hoping for is a world where every rule can be written down
in advance, and no interpretation of rules is ever necessary.
I suppose that since no one else has written a freeware distribution of
MS-DOS, the fact that your drivers dynamically link with MS-DOS means
that they are "one program", and you are therefore misappropriating
Microsloth's program by using the subterfuge of distributing drivers
separately from MS-DOS?
No, because neither MS-DOS nor the packet driver spec were written as
an escape from the GPL.
-russ <nel...@crynwr.com> http://www.crynwr.com/crynwr/nelson.html
Crynwr Software | Crynwr Software sells packet driver support | ask4 PGP key
11 Grant St. | +1 315 268 1925 (9201 FAX) | What is thee doing about it?
Potsdam, NY 13676 | LPF member - ask me about the harm software patents do.