Re: Yet another JDK1.1 llicence question
On Mon, Oct 21, 2002 at 09:45:59AM -0700, Stephen Zander wrote:
> >>>>> "Anthony" == Anthony Towns <firstname.lastname@example.org> writes:
> Anthony> Debian and denial are remarkably similar words. Quoting
> Anthony> mantras like that don't really further anyone's
> Anthony> understanding of anything. If you consider
> Anthony> stable/main/binary-i386/* to be a "product", then it's
> Anthony> completely fair and reasonable to say that
> Anthony> stable/non-free/binary-i386/* and hence the jdk is
> Anthony> "integrated" (/usr instead of /usr/local, postinst
> Anthony> scripts, etc), bundled (they're all under dists/stable)
> Anthony> and associated (they're on the same FTP site, their
> Anthony> dependencies are related, etc) with that product.
> OK, so the meaning of that phrase revolves around the definition of
> Product. If any exchange of value is implicit in the definition of
> Product then it doesn't apply. If a Product is created through simple
> aggregation, that phrase does apply and we can stop discussing things
> as Debian can't distribute the package. Any actual lawyers on this
It'd be a remarkably convoluted definition for "the Debian 3.0
distribution for i386" to not be a "product", IMO.
> >> The catch here is that Sun licenses the JDK & JRE slightly
> >> differently and Debian's policy doesn't deal well with
> >> sublicensing. I can Try and break the file in two if that
> >> would make things clearer.
> Anthony> We don't care about sublicensing; [...]
> You do care about sublicensing, Anthony. If I have two components in
> the same source tarball with different restriction on their use (the
> situation here), you have to care. What I didn't think Debian cared
> about, but apparently some parts of it do, are the legal restrictions
> under which I, as a Blackdowm team member independant of my Debian
> membership, operate.
We don't care about the details of what arrangements upstream makes,
except in so far that we have a license that's valid and allows us to
distribute the software, and that our users have (or that we can give
them) a license to use the software. It's possible we don't even care
that much about the latter actually.
We don't care what restrictions you operate under, we care about the
restrictions *we* operate under, and the copyright file needs to be
fairly clear about that. Having multiple licenses, some of which may
override sections of others, and some of which appear not to let us
distribute the software at all doesn't work for that.
> Anthony> The main thing is to make it clear what we can and can't
> Anthony> do.
> Yep. Hence the suplemental terms allowing for distribution indepndent
> of use. I'm still trying to establish what other conditions need to
> be addressed.
The main thing is to make things *clear*. If we've got our license C,
that Blackdown's required to have made similar to license S, then we only
care about C -- putting S in the copyright file as well is unnecessary
and confusing. If bits of S are needed, but bits are superseded, it
needs to be clear which bits are which. If some special exemptions are
being made for Debian that aren't being made for random Debian users,
that needs to be made clear too (and is quite okay for non-free).
If Blackdown aren't able to give Debian permission to distribute the
JDK/etc, and Sun haven't done so, we're obviously wasting time. If,
otoh, they are, then making it clear that that's okay (and what, if any,
conditions apply) is all that's necessary. If the software's binary
only (ie, Blackdown compile it, and Debian just packages it) a simple
statement that Debian can package and distribute copies of the compiled
binaries should be all that's needed and no major burden, although the
copyright file should also contain any terms appropriate for users who
actually want to use the package.
In summary: make sure upstream is aware of what we do, and that they
don't mind; then make it clear in the copyright file that that's the case.