Why-CLISP-is-under-GPL

From rms@gnu.ai.mit.edu Mon Oct 19 00:06:25 1992
Return-Path: <rms@gnu.ai.mit.edu>
Received: from mole.gnu.ai.mit.edu ([128.52.46.33]) by ma2s2.mathematik.uni-karlsruhe.de (4.1/SMI-4.0)
 id AA23967; Mon, 19 Oct 92 00:06:23 +0100
Received: by mole.gnu.ai.mit.edu (5.65/4.0)
 id <AA25733@mole.gnu.ai.mit.edu>; Sun, 18 Oct 92 18:59:05 -0400
Date: Sun, 18 Oct 92 18:59:05 -0400
From: rms@gnu.ai.mit.edu (Richard Stallman)
Message-Id: <9210182259.AA25733@mole.gnu.ai.mit.edu>
To: haible@ma2s2.mathematik.uni-karlsruhe.de
Subject: Common Lisp and Readline
Status: RO

I hate to have to play this role with a fellow hacker, but...

If you don't change to using the GPL, then you'll have to stop using
readline. Readline's terms say that the whole program has to be under
the GPL, and just having the user do the link doesn't change this. If
the program is designed to run with readline as a part, then readline
is a part of it.

An ordinary user can decide to link readline with a program and run
it. That's different--this user is not doing distribution, and thus
the GPL does not restrict what he can do.

From haible Fri Oct 30 18:54:38 1992
Return-Path: <haible>
Received: by ma2s2.mathematik.uni-karlsruhe.de (4.1/SMI-4.0)
 id AA28745; Fri, 30 Oct 92 18:54:37 +0100
Date: Fri, 30 Oct 92 18:54:37 +0100
From: haible (Bruno Haible)
Message-Id: <9210301754.AA28745@ma2s2.mathematik.uni-karlsruhe.de>
To: haible@ma2s2.mathematik.uni-karlsruhe.de, rms@gnu.ai.mit.edu
Subject: Re: Common Lisp and Readline
Status: RO


> I hate to have to play this role with a fellow hacker, but...

I'm sorry too, as I am very indebted to the GNU project.

> Readline's terms say that the whole program has to be under
> the GPL, and just having the user do the link doesn't change this.

I don't agree. My lisp.a is not a "work based on libreadline.a". What I
distribute is a "mere aggregation" of lisp.a and libreadline.a - the latter
with source.

> If the program is designed to run with readline as a part, then readline
> is a part of it.

I could provide a libnoreadline.a and let the user choose to link lisp.a
with either GNU's libreadline.a or my libnoreadline.a . Would that convince
you that lisp.a "can be reasonably considered independent and separate work" ?

Bruno Haible

From rms@gnu.ai.mit.edu Fri Oct 30 22:19:16 1992
Return-Path: <rms@gnu.ai.mit.edu>
Received: from mole.gnu.ai.mit.edu ([128.52.46.33]) by ma2s2.mathematik.uni-karlsruhe.de (4.1/SMI-4.0)
 id AA28939; Fri, 30 Oct 92 22:19:14 +0100
Received: by mole.gnu.ai.mit.edu (5.65/4.0)
 id <AA04907@mole.gnu.ai.mit.edu>; Fri, 30 Oct 92 16:09:54 -0500
Date: Fri, 30 Oct 92 16:09:54 -0500
From: rms@gnu.ai.mit.edu (Richard Stallman)
Message-Id: <9210302109.AA04907@mole.gnu.ai.mit.edu>
To: haible@ma2s2.mathematik.uni-karlsruhe.de
In-Reply-To: <9210301754.AA28745@ma2s2.mathematik.uni-karlsruhe.de> (haible@ma2s2.mathematik.uni-karlsruhe.de)
Subject: Common Lisp and Readline
Status: RO

 > Readline's terms say that the whole program has to be under
 > the GPL, and just having the user do the link doesn't change this.

 I don't agree. My lisp.a is not a "work based on libreadline.a". What I
 distribute is a "mere aggregation" of lisp.a and libreadline.a - the latter
 with source.

 I could provide a libnoreadline.a and let the user choose to link lisp.a
 with either GNU's libreadline.a or my libnoreadline.a . Would that convince
 you that lisp.a "can be reasonably considered independent and separate work" ?

No, and I doubt it would convince a judge either.

I'll have to insist that you stop using readline unless you make the
program free.


From haible Sat Oct 31 01:08:00 1992
Return-Path: <haible>
Received: by ma2s2.mathematik.uni-karlsruhe.de (4.1/SMI-4.0)
 id AA29062; Sat, 31 Oct 92 01:08:00 +0100
Date: Sat, 31 Oct 92 01:08:00 +0100
From: haible (Bruno Haible)
Message-Id: <9210310008.AA29062@ma2s2.mathematik.uni-karlsruhe.de>
To: haible@ma2s2.mathematik.uni-karlsruhe.de, rms@gnu.ai.mit.edu
Subject: Re: Common Lisp and Readline
Status: RO


> I'll have to insist that you stop using readline unless you make the
> program free.

I built a libnoreadline.a that can be linked together with lisp.a,
replacing libreadline.a .

I will reorganize the distribution into 2 independent parts:
* clisp.lzh containing lisp.a and libnoreadline.a,
* readline.tar.Z containing libreadline.a and its source.

The first one is enough to build a CLISP executable. It contains
no GNU parts.

> I doubt it would convince a judge either.

The only thing CLISP will have to do with the readline library
is that *THE USER* *MAY OPTIONALLY* link CLISP with the readline library.

No judge will admit that this gives you the right to determine the copyright
of CLISP.

>From your own words:
> An ordinary user can decide to link readline with a program and run
> it. That's different--this user is not doing distribution, and thus
> the GPL does not restrict what he can do.


If you still insist on me being wrong, then please tell me what part
of the GNU GPL I am breaking. Or update the GNU GPL.

Bruno Haible

From haible Sat Oct 31 01:08:26 1992
Return-Path: <haible>
Received: by ma2s2.mathematik.uni-karlsruhe.de (4.1/SMI-4.0)
 id AA29068; Sat, 31 Oct 92 01:08:25 +0100
Date: Sat, 31 Oct 92 01:08:25 +0100
From: haible (Bruno Haible)
Message-Id: <9210310008.AA29068@ma2s2.mathematik.uni-karlsruhe.de>
To: haible@ma2s2.mathematik.uni-karlsruhe.de, rms@gnu.ai.mit.edu
Subject: Readline & GPL
Status: RO


If someone was to turn the readline library into a shared library
(a library that is needed at runtime by the executable, but not
contained in the executable):
Would that mean that any executable that uses a readline
shared lib would have to be accompanied with full source? Or would it, in the
contrary, mean that such an executable may use the readline lib without
"containing" it and without being affected by such copyright issues?

Bruno Haible

From rms@gnu.ai.mit.edu Sat Oct 31 01:11:59 1992
Return-Path: <rms@gnu.ai.mit.edu>
Received: from mole.gnu.ai.mit.edu ([128.52.46.33]) by ma2s2.mathematik.uni-karlsruhe.de (4.1/SMI-4.0)
 id AA29075; Sat, 31 Oct 92 01:11:56 +0100
Received: by mole.gnu.ai.mit.edu (5.65/4.0)
 id <AA06772@mole.gnu.ai.mit.edu>; Fri, 30 Oct 92 19:02:38 -0500
Date: Fri, 30 Oct 92 19:02:38 -0500
From: rms@gnu.ai.mit.edu (Richard Stallman)
Message-Id: <9210310002.AA06772@mole.gnu.ai.mit.edu>
To: haible@ma2s2.mathematik.uni-karlsruhe.de
In-Reply-To: <9210310008.AA29068@ma2s2.mathematik.uni-karlsruhe.de> (haible@ma2s2.mathematik.uni-karlsruhe.de)
Subject: Readline & GPL
Status: RO

 If someone was to turn the readline library into a shared library
 (a library that is needed at runtime by the executable, but not
 contained in the executable):
 Would that mean that any executable that uses a readline
 shared lib would have to be accompanied with full source?

Yes.

From rms@gnu.ai.mit.edu Sat Oct 31 01:29:01 1992
Return-Path: <rms@gnu.ai.mit.edu>
Received: from mole.gnu.ai.mit.edu ([128.52.46.33]) by ma2s2.mathematik.uni-karlsruhe.de (4.1/SMI-4.0)
 id AA29083; Sat, 31 Oct 92 01:28:59 +0100
Received: by mole.gnu.ai.mit.edu (5.65/4.0)
 id <AA06922@mole.gnu.ai.mit.edu>; Fri, 30 Oct 92 19:19:36 -0500
Date: Fri, 30 Oct 92 19:19:36 -0500
From: rms@gnu.ai.mit.edu (Richard Stallman)
Message-Id: <9210310019.AA06922@mole.gnu.ai.mit.edu>
To: haible@ma2s2.mathematik.uni-karlsruhe.de
In-Reply-To: <9210310008.AA29062@ma2s2.mathematik.uni-karlsruhe.de> (haible@ma2s2.mathematik.uni-karlsruhe.de)
Subject: Common Lisp and Readline
Status: RO

 I built a libnoreadline.a that can be linked together with lisp.a,
 replacing libreadline.a .

 I will reorganize the distribution into 2 independent parts:
 * clisp.lzh containing lisp.a and libnoreadline.a,
 * readline.tar.Z containing libreadline.a and its source.

 The first one is enough to build a CLISP executable. It contains
 no GNU parts.

True. If that were the whole situation--if readline did not exist--
then I would have no grounds to object.

However, the sum total of what you are doing is still tantamount to
distributing one program which contains readline but is not under the
GPL.

This is a problem I cannot ignore. If you can get away with this then
any company can get away with it. The result would be to negate for
practical purposes the GPL's requirement that improvements be free.
And that would deprive GNU software of a major source of improvements.

Is that what you want?



From rms@gnu.ai.mit.edu Fri Nov 6 11:02:40 1992
Return-Path: <rms@gnu.ai.mit.edu>
Received: from mole.gnu.ai.mit.edu ([128.52.46.33]) by ma2s2.mathematik.uni-karlsruhe.de (4.1/SMI-4.0)
 id AA13962; Fri, 6 Nov 92 11:02:36 +0100
Received: by mole.gnu.ai.mit.edu (5.65/4.0)
 id <AA09628@mole.gnu.ai.mit.edu>; Fri, 6 Nov 92 04:52:01 -0500
Date: Fri, 6 Nov 92 04:52:01 -0500
From: rms@gnu.ai.mit.edu (Richard Stallman)
Message-Id: <9211060952.AA09628@mole.gnu.ai.mit.edu>
To: haible@ma2s2.mathematik.uni-karlsruhe.de
In-Reply-To: <9210310008.AA29062@ma2s2.mathematik.uni-karlsruhe.de> (haible@ma2s2.mathematik.uni-karlsruhe.de)
Subject: Common Lisp and Readline
Status: RO

If you do succeed in circumventing the GPL for readline, you would be
blazing a path for every commercial company that wants to do it.
Would you really like that result?

Wouldn't it be best to eliminate this dispute by using the GPL for
Common Lisp? Totally aside from readline, the Lisp system would be
more useful that way. For example, maybe it could be a useful part of
GNU.



From haible Fri Nov 6 16:13:07 1992
Return-Path: <haible>
Received: by ma2s2.mathematik.uni-karlsruhe.de (4.1/SMI-4.0)
 id AA14116; Fri, 6 Nov 92 16:13:06 +0100
Date: Fri, 6 Nov 92 16:13:06 +0100
From: haible (Bruno Haible)
Message-Id: <9211061513.AA14116@ma2s2.mathematik.uni-karlsruhe.de>
To: haible@ma2s2.mathematik.uni-karlsruhe.de, rms@gnu.ai.mit.edu
Subject: Common Lisp, Readline and GPL
Status: RO



> I will reorganize the distribution into 2 independent parts:
> * clisp.lzh containing lisp.a and libnoreadline.a,
> * readline.tar.Z containing libreadline.a and its source.
>
> The first one is enough to build a CLISP executable. It contains
> no GNU parts.
>
> However, the sum total of what you are doing is still tantamount to
> distributing one program which contains readline but is not under the
> GPL.

Not totally. The sum of what I AND THE USER are doing is tantamount to
building a program which contains readline but is not under the GPL.
But
 * the user isn't distributing anything,
 * I am not responsible for the user's deeds,
 * I am not distributing "one program", so GPL doesn't apply to me either.


> If that were the whole situation--if readline did not exist--
> then I would have no grounds to object.

So the only reason why you may object is that my lisp.a has an INTERFACE
that allows it to be linked to libreadline.a . As far as I know, only
the readline CODE is copyrighted, not its INTERFACE. (Would you object
because my libnoreadline.a has a similar interface?)


> This is a problem I cannot ignore. If you can get away with this then
> any company can get away with it.
> ...
> And that would deprive GNU software of a major source of improvements.
> Is that what you want?

No, of course! But I would like to see the problem analyzed in depth.

Problem 1: If someone (like me) lets the user do the final link phase.

Problem 2: If the final link phase is done by the OS automatically at run time
 (shared libraries).

You should update GPL for these cases.


> If you do succeed in circumventing the GPL for readline, you would be
> blazing a path for every commercial company that wants to do it.

But the problems remain there even if I put CLISP under GPL.


From haible Fri Nov 6 16:13:19 1992
Return-Path: <haible>
Received: by ma2s2.mathematik.uni-karlsruhe.de (4.1/SMI-4.0)
 id AA14123; Fri, 6 Nov 92 16:13:18 +0100
Date: Fri, 6 Nov 92 16:13:18 +0100
From: haible (Bruno Haible)
Message-Id: <9211061513.AA14123@ma2s2.mathematik.uni-karlsruhe.de>
To: haible@ma2s2.mathematik.uni-karlsruhe.de, rms@gnu.ai.mit.edu
Subject: Common Lisp, Readline and GPL
Status: RO


> Wouldn't it be best to eliminate this dispute by using the GPL for
> Common Lisp?

Would be best for me, true.

> Totally aside from readline, the Lisp system would be
> more useful that way.

This and the following are convincing me:

* Up to 1991 the decision whether using GPL or not, was simply a matter of
 philosophy or copyright policy.
 The success of Linux, however, demonstrates how a speed and quality of
 development was achieved which would have been impossible without access
 for everyone to the source of everything.

* Finding co-developers for other Lisp packages or testers for other
 hardware/OS platforms might be easier if I release full source.

So be prepared to seeing CLISP's source before Christmas.

Yours,
 B. Haible

From rms@gnu.ai.mit.edu Fri Nov 6 21:31:33 1992
Return-Path: <rms@gnu.ai.mit.edu>
Received: from mole.gnu.ai.mit.edu ([128.52.46.33]) by ma2s2.mathematik.uni-karlsruhe.de (4.1/SMI-4.0)
 id AA15204; Fri, 6 Nov 92 21:31:28 +0100
Received: by mole.gnu.ai.mit.edu (5.65/4.0)
 id <AA10933@mole.gnu.ai.mit.edu>; Fri, 6 Nov 92 15:20:48 -0500
Date: Fri, 6 Nov 92 15:20:48 -0500
From: rms@gnu.ai.mit.edu (Richard Stallman)
Message-Id: <9211062020.AA10933@mole.gnu.ai.mit.edu>
To: haible@ma2s2.mathematik.uni-karlsruhe.de
Cc: rms@gnu.ai.mit.edu
In-Reply-To: <9211061513.AA14116@ma2s2.mathematik.uni-karlsruhe.de> (haible@ma2s2.mathematik.uni-karlsruhe.de)
Subject: Common Lisp, Readline and GPL
Status: RO

 * the user isn't distributing anything,
 * I am not responsible for the user's deeds,
 * I am not distributing "one program", so GPL doesn't apply to me either.

The FSF position would be that this is still one program, which has
only been disguised as two. The reason it is still one program is
that the one part clearly shows the intention for incorporation of the
other part.


I say this based on discussions I had with our lawyer long ago. The
issue first arose when NeXT proposed to distribute a modified GCC in
two parts and let the user link them. Jobs asked me whether this was
lawful. It seemed to me at the time that it was, following reasoning
like what you are using; but since the result was very undesirable for
free software, I said I would have to ask the lawyer.

What the lawyer said surprised me; he said that judges would consider
such schemes to be "subterfuges" and would be very harsh toward
them. He said a judge would ask whether it is "really" one program,
rather than how it is labeled.

So I went back to Jobs and said we believed his plan was not allowed
by the GPL.

The direct result of this is that we now have an Objective C front
end. They had wanted to distribute the Objective C parser as a
separate proprietary package to link with the GCC back end, but since
I didn't agree this was allowed, they made it free.


So I don't think the GPL actually requires a correction for this.
But perhaps it would be a good idea to add a note explaining this.


From haible Sun Nov 8 16:31:52 1992
Return-Path: <haible>
Received: by ma2s2.mathematik.uni-karlsruhe.de (4.1/SMI-4.0)
 id AA18675; Sun, 8 Nov 92 16:31:51 +0100
Date: Sun, 8 Nov 92 16:31:51 +0100
From: haible (Bruno Haible)
Message-Id: <9211081531.AA18675@ma2s2.mathematik.uni-karlsruhe.de>
To: haible@ma2s2.mathematik.uni-karlsruhe.de, rms@gnu.ai.mit.edu
Subject: Re: Common Lisp, Readline and GPL
Status: RO


Thanks you for the explanation of the lawyers' point of view.

> What the lawyer said surprised me; he said that judges would consider
> such schemes to be "subterfuges" and would be very harsh toward
> them. He said a judge would ask whether it is "really" one program,
> rather than how it is labeled.

I accept this.

> They had wanted to distribute the Objective C parser as a
> separate proprietary package to link with the GCC back end, but since
> I didn't agree this was allowed, they made it free.

The situation for CLISP is a bit different: CLISP was running without the
readline library for over one year before I began using readline() in
September 1992. That was also the reason why I was able to build a drop-in
replacement (libnoreadline.a) within only one hour.

The readline library is not an essential or integral part of CLISP - it makes
the user interface nicer, and that's all.


From rms@gnu.ai.mit.edu Mon Oct 11 18:57:43 1993
Return-Path: <rms@gnu.ai.mit.edu>
Received: from mole.gnu.ai.mit.edu by ma2s2.mathematik.uni-karlsruhe.de (4.1/SMI-4.0)
 id AA06403; Mon, 11 Oct 93 18:57:38 +0100
Received: by mole.gnu.ai.mit.edu (5.65/4.0)
 id <AA09632@mole.gnu.ai.mit.edu>; Mon, 11 Oct 93 13:56:03 -0400
Date: Mon, 11 Oct 93 13:56:03 -0400
From: rms@gnu.ai.mit.edu (Richard Stallman)
Message-Id: <9310111756.AA09632@mole.gnu.ai.mit.edu>
To: haible@ma2s2.mathematik.uni-karlsruhe.de
In-Reply-To: <9310111623.AA06248@ma2s2.mathematik.uni-karlsruhe.de> (haible@ma2s2.mathematik.uni-karlsruhe.de)
Subject: Re: compiling with gcc without linking with libggc.a
Status: RO

 But libgmp.a should remain under GPL.

I'm interested in hearing why you think so.
Someone else is arguing that it should be under the LGPL;
I haven't made up my mind yet.

From haible Tue Oct 12 13:28:37 1993
Return-Path: <haible>
Received: by ma2s2.mathematik.uni-karlsruhe.de (4.1/SMI-4.0)
 id AA07494; Tue, 12 Oct 93 13:28:36 +0100
Date: Tue, 12 Oct 93 13:28:36 +0100
From: haible (Bruno Haible)
Message-Id: <9310121228.AA07494@ma2s2.mathematik.uni-karlsruhe.de>
To: rms@gnu.ai.mit.edu
Subject: Re: libgmp.a
Status: RO

> But libgmp.a should remain under GPL.
> I'm interested in hearing why you think so.

Because one can build a multi-precision calculator (like `bc') as a wrapper
to libgmp.a. Capability of handling arbitrary size numbers is still an
essential advantage (and marketing issue) for computer algebra systems and
Lisp implementations.

I had to put CLISP under GPL because libreadline.a was something I wanted
to use and I didn't want to write myself. libgmp.a is similar in this
respect, and it may therefore help putting some new packages under GPL.

Building libgmp.a was just too hard work. Other people shouldn't get it for
free.

Bruno Haible

From rms@gnu.ai.mit.edu Tue Oct 12 18:28:27 1993
Return-Path: <rms@gnu.ai.mit.edu>
Received: from mole.gnu.ai.mit.edu by ma2s2.mathematik.uni-karlsruhe.de (4.1/SMI-4.0)
 id AA08120; Tue, 12 Oct 93 18:28:25 +0100
Received: by mole.gnu.ai.mit.edu (5.65/4.0)
 id <AA14488@mole.gnu.ai.mit.edu>; Tue, 12 Oct 93 13:26:57 -0400
Date: Tue, 12 Oct 93 13:26:57 -0400
From: rms@gnu.ai.mit.edu (Richard Stallman)
Message-Id: <9310121726.AA14488@mole.gnu.ai.mit.edu>
To: haible@ma2s2.mathematik.uni-karlsruhe.de
In-Reply-To: <9310121228.AA07494@ma2s2.mathematik.uni-karlsruhe.de> (haible@ma2s2.mathematik.uni-karlsruhe.de)
Subject: Re: libgmp.a
Status: RO

Thanks for reminding me about this example.
I agree it is a strong argument.