Re: Eclipse 3.0 Running ILLEGALY on Kaffe
> Raul Miller <email@example.com> 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?
On Sat, Jan 15, 2005 at 10:40:08PM -0500, Brian Thomas Sniffen wrote:
> Copies of his work -- of Kaffe, in this case.
Kaffe is GPLed. I can distribute Kaffe unless I violate the terms of
So it's up to you to show that I would be violating the GPL when I
distribute both Eclipse 3.0 and Kaffe...
> > 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".
Indeed it does. So what's your basis for saying that "Eclipse 3.0 and
Kaffe" is "a work"? Is it a "work based on the Program"? If it is,
then which of the following is it:
[a] the Program? (obviously not -- there are no GPL notices on Eclipse
[b] any derivative work under copyright law?
I'm guessing you're thinking that the combination constitutes a
derivative work under copyright law. But what is your basis for this?
This is not a rhetorical question, I'm asking for a specific cite of
some relevant law or legal precedent.
But that's a pretty general request, so I'll try to point out, as clearly
as I can, what the overall issues are and what I think are the significant
points within those issues -- in other words I plan on outlining in this
message the things which I'd want to make sure are addressed before I
considered a legal cite relevant.
In other words, I'm asking "why does eclipse 3.0 + kaffe constitute a
derived work of kaffe under copyright law"?
> 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.
Only if Eclipse 3.0 + Kaffe constitutes a derived work. But how can
that be, when the total content which distinguishes this "work" from the
"non-work" is less than 80 bytes of text?
(I'm guessing that any argument that you can advance about why this must
be a derived work must treat the DFSG as terms enforced by the GPL, or is
based on the idea that copyright protects process or ... I guess I should
let you explain your reasoning. If only I could get you to do so...)
> > 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.
So what part of copyright law makes "Eclipse 3.0 + Kaffe" a copyrightable
work, other than the fact that they're on the same volume?
What copyrightable work was done to make this combination happen?
Remember, we already have Eclipse 3.0 as a running program without Kaffe.
We don't have it in main, but your argument would keep the combination
out of non-free as well as main.
> > 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.
But is that enough evidence? For example, depending on how the user has
put together their system, eclipse might very well not invoke Kaffe.
(Which, combined with basic information about java, is sufficient to
show that "eclipse works with kaffe" is not the result of any specific
work to combine the two, but instead is the result of both following
the same standards.)
> 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.
You're ignoring Kaffe's explicit permission, stating this kind of thing
is ok even if someone might think the GPL says otherwise. But ok,
let's focus on proof that the combination is not mere aggregation.
Copyright law protects works of art, science, etc. Copyright does not
protect ideas, procedures, methods of operation or mathematical concepts
As near as I can tell, the copyrightable work consists solely of:
[a] Depends: kaffe | ...
[b] Build-Depends: kaffe | ...
[c] The work of choosing both kaffe and eclipse to go onto the machines
of Debian's users.
[d] The program runs.
[e] The association of eclipse -> jvm -> kaffe which will exist on some
[a] and [b] constitute significantly less than a sentence, so aren't
copyrightable -- they're too trivial to count in and of themselves.
[c] is covered under mere aggregation.
[d] and [e] not protected by copyright law.
In the more general case, [d] and [e] would indicate that more work
than mere aggregation had taken place to form the derived work. But we
already had Eclipse and Kaffe. We were just preventing them from being
installed in a fashion where eclipse would work with Kaffe.
But that distinction -- preventing from running vs. running -- is not
in and of itself a concern for copyright law.
> 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.
How about eclipse + kaffe, where we've two words different such that
eclipse won't invoke kaffe? Because if it's illegal with those two
words it has to be illegal without those two words.
> > 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.
If we do, then just getting rid of the Depend: and Build-Depends:
won't be enough. Those phrases aren't sufficient content to matter for
And, just to repeat myself, copyright law does not concern itself with
what the programs do. What the programs do is just evidence that might
be used in court to show what kind of work was done.
If this were a C program -- where portable code requires significant
work, the fact that "it runs" would probably be sufficient evidence
to show that there's some art involved. But this is a Java program,
so as a general rule, we will need more evidence than "it runs".