[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: IBM doesn't like the GPL

From: Alexander Terekhov
Subject: Re: IBM doesn't like the GPL
Date: Mon, 23 Mar 2009 18:56:31 +0100

Hyman Rosen wrote:
> GPLv3 has cleaned up the language:
>      <>

GPLv3 is clear as mud.

Outside the Definitions section it purports to define an "is called"
term "aggregate"

"A compilation of a covered work with other separate and independent
works, which are not by their nature extensions of the covered work, and
which are not combined with it such as to form a larger program, in or
on a volume of a storage or distribution medium, is called an
“aggregate” if the compilation and its resulting copyright are not used
to limit the access or legal rights of the compilation's users beyond
what the individual works permit."

(in turn using undefined terms "extensions" and "larger program")

and then postulates that

"Inclusion of a covered work in an aggregate does not cause this License
to apply to the other parts of the aggregate."

"This paragraph seems to generate more legal uncertainty, something that
has not been lost on commentators of the draft.34 This paragraph tries
to rationalise specific cases in which the revamped GPL copyleft clause
will apply by inventing a new definition to what a compilation is. They
try to determine that the software distributed in the same distribution
medium has to be GPL if it is a "compilation", but not if it is an
"aggregate". Why create the new terminology?"'CF.NoteStartNodeId'%20LIKE%20'gpl3.distribmod.p5.s2'%20AND%20'CF.NoteUrl'%20LIKE%20'gplv3-draft-4'%20

Comment 3380: Combining propietary and GPL code 
This Comment is resolved by:

This Comment is part of the discussion on:

Regarding the text: Inclusion of a covered work in an aggregate does not
cause this License to apply to the other parts of the aggregate
In section: gpl3.distribmod.p5.s2
Submitted by: user on 2007-06-15 at 01:47 EDT
0 agree: 
noted by user on 2007-06-15 at 01:47 EDT: 

Lawrence Rosen has stated that this allows the combination of propietary
code and GPLV3 code. Now he may be absolutely and totally wrong on that
matter but his statement casts doubt on this provision. It is aboslutely
paramount that there is absolutely no doubt about the fact that you can
not combine propietary and GPL(V3) code at all. That goes against the
central design choice of the GNU GPL as a copyleft license. There must
be ABSOLUTELY no doubt at all, among ANY reasonably thinking human being
about that, at all. Period.

This needs to be reqorded so that there cannot be anyt such doubt what
so ever.

noted by veatnik on 2007-06-27 at 16:20 EDT: 

False. Many people build GPL applications on top of systems with
proprietary versions of system libraries (linking) and distribute the
code and or binaries. Hopefully the libraries have well documented APIs
so that GPL libraries could be made if desired. If what you say is true
then we currently do not follow this in practice or no GPLed programs
would run on Windows. (This would be no loss to me but I think it is
reasonable to allow a GPLed program to run on Windows.) This does also
point to the issue that comes legally with some binary modules. What do
we do with a legally required binary module for using a wireless card.
FCC says we can't tamper with the module that ensures FCC compliance. If
we want to write a GPL driver for the wireless card then we need to link
to the binary module. If the API for using the module is fully
documented so that an end user can effectively fix the driver then I do
not see a problem with that and it seems to be allowed by the
"aggregated" clause BUT...the clause does not insist that interface
points in an aggregation must be documented and I think that this should
be required. I also think that we should harmonize the LGPL into this so
that object modules must be provided for relinking a version of an
aggregation. I suspect that Lawrence Rosen is exactly right and the
history and common practice of using GPL code agrees and sets the
foundation in court for the GPL to be ruled weaker than the LGPL in its

As this is not the intent of many here we should bring the GPL up to the
same standard that is given in the LGPL and fix the binary module issue
once and for all or we will keep seeing issues with this.

(Actually nothing really prevents a user from coding and making a
replacement for a binary module if they have enough knowledge about
internal requirements. Although there are potential legal pitfalls if
the result is illegal. In other words as long as the result meets all
applicable laws then we should not preclude someone from replacing a
module. Another way to say that objects should be provided for linking.)

More broadly to my satisfaction, though, the GPLv3 license contains
clearer language that will encourage its software to be used in
“aggregations” (what in United States law we call “compilations” or
“collective works”) that include many independent works. Computer
professionals will better understand these independent works as modules
of code, so I’ll use that word instead. As long as those modules are
independent works under copyright law, and as long as their own
independent licenses allow this to be done, verbatim copies of those
modules can be combined to build larger systems of increasing
sophistication and breadth. GPLv3, as I now read Draft 4 the license,
allows that. 

I have long counseled that GPLv2 allowed that result also, but some
argued otherwise. All the FUD that I and others have spread over the
years about linking—static, dynamic, otherwise—didn’t help matters much.
This argument can be replaced now by intelligent conversation about the
actual words of GPLv3, which I understand to allow aggregations of
verbatim copies of independently-written modules taken from the
worldwide portfolio of free and open source software. I have always read
the Open Source Definition as mandating that open source licenses permit
the copying and distribution of verbatim copies including for collective
works, and this added clarity in GPLv3 supports that interpretation.
Perhaps now our software commons is vastly larger for collecting
independent open source modules into interoperating solutions for our

Subject: RE: Comments on GPLv3 Actions... 
From: Lawrence Rosen ( 
Date: Jun 6, 2007 3:12:33 pm 

At Niclas Hedman's request, I'm copying legal-discuss. 

-----Original Message-----
From: [] On Behalf Of Niclas
Sent: Tuesday, June 05, 2007 9:27 PM
Subject: Re: Comments on GPLv3

Does this mean that the virality of Collective Works is gone?

1) Copyleft for derivative works remains.
2) Copyleft for collective works never was.

The skill comes in learning how to distinguish them. GPLv3 uses clearer
language and makes the distinctions more obvious than GPLv2. In
particular, "verbatim copies" may be used in "aggregations." 
Ultimately, though, copyright law defines those terms.

Ok. But how does that relates to the "Derivative Works" definition of
dynamically linked languages, such as Java?? IIUIC, FSF always
maintained that the runtime linked result is always a derivative
works. More importantly, got any info on how this affects LGPL ?? Or 
is it not part of v3 at all?

On list would be much better. I know a lot of people
are following this closely.

[LR:] The definitions of derivative works and collective works are in
the Copyright Act, and are not specific to dynamically linked 
languages such as Java. 

The law deals with copyrighted works. When work A is aggregated with
work B to form a collective work C, the law doesn't ask about the 
type of glue that is used in the book binding to collect those works 
together. An expressive work that contains a data base module A and a 
file system module B can be collected into a larger program C that 
serves as a software application. Copyright subsists in C to the 
extent that the aggregation is itself an expressive work. As long as 
the individual licenses for A and B allow the aggregation and 
distribution of verbatim copies (and all FOSS licenses do!),
then it matters little whether the aggregation is via Java mechanisms 
or old-style linking in C. It makes no difference if the resulting 
work is dynamically loaded or pre-etched onto computer chips. 

To call every such aggregation a derivative work misapplies the
copyright law. Independently-written software modules are not 
derivative works of each other simply because they are copied, 
verbatim, into an operating computer and operated together. In fact, 
for one copyrighted work to attempt to control (through its license) 
the licenses of independent works may be, in some situations, 
copyright misuse resulting in the loss of the copyright. 

The only thing new that GPLv3 brings to the table about this point is
that its language is now clearer than GPLv2. Verbatim copies of
GPLv3-licensed works can be made and distributed without any effect 
on the licenses of aggregated works. (See GPLv3 §4 and last 
paragraph of §5.) The GPLv3 license continues to apply to those 
verbatim copies.

I have not focused at all on LGPLv3 and so I have no comment on it at
this time. But it certainly will be interesting to see if LGPLv3 
attempts to do more or less than what GPLv3 does for aggregations of 
verbatim copies. 


P.S. Be aware that RMS doesn't agree with me. I'm not surprised.

DISCLAIMER: Discussions on this list are informational and educational
only.  Statements made on this list are not privileged, do not
constitute legal advice, and do not necessarily reflect the opinions
and policies of the ASF.  See <> for
official ASF policies and documents.
To unsubscribe, e-mail:
For additional commands, e-mail:


(GNG is a derecursive recursive derecursion which pwns GNU since it can
be infinitely looped as GNGNGNGNG...NGNGNG... and can be said backwards
too, whereas GNU cannot.)

reply via email to

[Prev in Thread] Current Thread [Next in Thread]