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

Re: Eclipse 3.0 Running ILLEGALY on Kaffe

Raul Miller <moth@debian.org> writes:

> On Sat, Jan 15, 2005 at 02:31:13PM -0500, Brian Thomas Sniffen wrote:
>> Again, this isn't about the copyright holder's right to control
>> production of derived works.  This is about the copyright holder's
>> right to control copying and distribution of copies.  Reading GPL 2b,
>> I cannot see permission to distribute a CD with Eclipse and Kaffe on
>> it, such that Eclipse runs on top of Kaffe when I insert the CD.
> Copies of what?

Copies of his work -- of Kaffe, in this case.

> The license on Kaffe applies to Kaffe and ...
>   "This License applies to any program or other work which contains a
>    notice placed by the copyright holder saying it may be distributed
>    under the terms of this General Public License. The "Program", below,
>    refers to any such program or work, and a "work based on the Program"
>    means either the Program or any derivative work under copyright law:
>    that is to say, a work containing the Program or a portion of it,
>    either verbatim or with modifications and/or translated into another
>    language. ..."
> If Kaffe + Eclipse 3.0 is not a derivative work of Kaffe under copyright
> law, then the rest of the GPL is irrelevant.

It says, and you quoted: "a work containing the Program or a portion
of it."  If with modifications, that's a derivative work.  If without
modifications that's "the Program" in the phrase "the Program or any
derivative work".

The license on Kaffe does not in any way inhibit distribution of
copies of Eclipse.  I don't believe for a second that Eclipse is
derivative of any particular JVM.  But Eclipse+Kaffe does contain a
copy of Kaffe.  The GPL grants permission for distribution of copies
of Kaffe.  It does this in its section 2.

GPL 2b says that if distributing a combined work which contains a copy
of a GPL'd work, then the entire result must be under the terms of the
GPL.  This is that case exactly.

> We have some evidence that Kaffe + Eclipse 3.0 might be a derivative
> work under copyright law (Eclipse runs with Kaffe as the JVM).
> We have some other evidence that Keffe + Eclipse 3.0 is not a derivative
> work under copyright law (we have people saying that Eclipse 3.0 uses
> only java byte codes whose semantics are independent of the JVM).
> We also have the evidence of the dependencies in Eclipse, which do not
> require Kaffe at any point -- Eclipse can be built and used just as
> easily with any of several JVMs.

To say it again clearly: I'm not talking about distribution of Eclipse
alone, but of Eclipse and Kaffe intertwined.

> None of these statements, in and of themselves, would settle the issue
> in the general case, but let's focus on this case.
> In other words: unless we have some very good evidence that Kaffe pluse
> Exlipse 3.0 is a "Program" in the sense defined by the GPL, the rest of
> the GPL requirements are irrelevant.

That doesn't have to be a Program.  It's enough that Kaffe is a
Program, and Eclipse+Kaffe is a "work that you distribute...in
whole...contains...the Program" (GPL 2b), and that it doesn't qualify
for the "mere aggregation" exception.

> Can you provide such evidence?

I can provide evidence that it doesn't qualify as mere aggregation --
the dependency and automated invocation of Kaffe by running "eclipse"
at a command line.

It seems plainly evident to me that the Debian OS is a work containing
the Program (Kaffe), so all the parts of Debian that aren't merely
aggregated with Kaffe have to be distributed under the GPL.  Emacs,
for example, is clearly merely aggregated.

It's easier for me to think about this when considering stripped down
distributions; a distribution of Emacs+Kaffe wouldn't be anything but
mere aggregation, but Eclipse+Kaffe would be integrated, more than
just aggregation.

> If not, do you have any suggestions about good ways to find such evidence?
> If not, what grounds do you have for you claiming that those GPL
> restrictions are relevant to this discussion?

I think that's pretty clear evidence that Debian does have to worry
about this.


Brian Sniffen                                       bts@alum.mit.edu

Reply to: