Re: testing before a release

[Etienne M. Gagnon]

Hi all,

I am surprized to see the direction that this licensing discussion is taking. 
No body has yet addressed the CURRENT PROBLEM of the license of non-AWT code.

Let me refresh your memory:  The current Classpath exception to the GPL allows 
proprietary code to include derivative work, WITHOUT any obligation to 
redistribute the Classpath source code, (and improvements made to it).  It 
doesn't even force the distributor of the derived proprietary work to display 
some acknowledgement of the presence of Classpath code.

I have made the following proposal, which seems to have fallen into oblivion: 
License Classpath under the LGPL, as this license carries the intent of most (if 
not all) Classpath contributors, AND add an exception to the LGPL to eliminate 
the requirement that forces dynamic linking or providing relinkable modules.

As for the AWT, you can go with the plain GPL if you want, but this is, in my 
humble opinion, a dead-end.  There's already other free software projects under 
licenses like the LGPL that provide things like "GTK" wrappers for Java.  (See 
the following project: http://java-gnome.sourceforge.net ).  Such code could 
easily serve as the basis to rewrite a new AWT under a much kinder license for 
applications.

I have serious reserves on an interpretation of the GPL that would say that it 
is ok to run a graphical proprietary application over a GPL'ed AWT.  I also have 
the concern that the GPL on the AWT would impose itself on the underlying VM, 
which in turn would impose the GPL on linked JNI modules.  This would invalidate 
my VM's exception for linking proprietary JNI code (and running proprietary 
applications).  It's your choice: you're the copyright holders, but nobody has 
to use your code and help you improve it, either.

We all know that writing a complete class library for Java is a major 
undertaking; I don't see why you want to do the GNU project (Classpath/libgcj) 
in a way that will cause duplicate efforts, because of unacceptable licenses, 
and code contribution policies (copyright assignment).

As usual, you can continue to ignore my suggestions; I'm have now got used to 
it.  Some of your leaders have their (partly hidden) political/commercial 
agendas.  This is why there's such a license mess.  Nobody cared about the LGPL 
license for Classpath before the libgcj merge.  To solve the "linking" problem, 
they went with GPL + exception, but a careful inspection of it tell you that it 
is a permission to reuse contributed code without any sourc-code redistribution 
obkigation (very good for companies [I could say names here] to make money $$$ 
selling embedded stuff that was developped by others).

A note about copyright assignment: You know, one reason the Linux kernel is so 
successful is that: even Linus CANNOT sell the code under a proprietary license, 
because HE IS NOT the sole copyright holder.  Because of this, the Linux kernel 
is de-facto a really Free piece of code. Nobody single person can take a 
proprietary advantage of it, unless ALL contributors agree.

Etienne

Mark Wielaard wrote:

> Hi,
>
> On Wed, Nov 07, 2001 at 08:31:55AM -0500, Brian Jones wrote:
>
> > Mark Wielaard <address@hidden> writes:
> >
> >
> > > I think this means one of two things. Either we wait till Per Bothner
> > > reports on the progress he has made for the gcc/libgcj project. Hopefully
> > > they come to some sort of agreement that can also be used for the Classpath
> > > project (having both Classpath AWT and libgcj AWT under the same license
> > > would be a big win). Or we make sure that all primary authors of our AWT
> > > implementation don't mind the license change to GPL and we change the
> > > copyright notice for all java.awt.* packages. If the primary authors
> > > (Aaron Renn, Brian Jones and Paul Fisher) mail me this is OK then I will
> > > change the copyright notice as soon as possible.
> > To make a release, we need to change the java.awt.* license headers to
> > indicate the code is covered by the GPL.  We're doing this because as
> > the decision maker for our project, RMS has requested it and at this
> > time not doing so would be counterproductive.
> >
> > I'm attaching a emacs lisp script Paul wrote a long time ago to change
> > the license stuff for Japhar and it could be useful to someone now
> > who'd like to help by making this change.  It does look like a simple
> > perl or sed script could remove the "Library" part and that would be
> > enough.  Are you still volunteering to do this Mark?
>
> Yes, but still on the condition stated above, that the primary authors
> of the (gnu.)java.awt.* packages agree that this is the correct thing to
> do. I suppose you and Paul discussed this and agreed on it. That means
> that I would like to have Aaron his OK. Am I missing someone (the mailing
> list archives suggest that Jim Blair did a lot of work on this, but I
> don't know if much of his code is now in the tree and will be relicensed).
> If anybody else who did actual work on these packages objects please speak
> up now.
>
> We also need a LICENSE.README file precisely stating the impact of this
> relicensing. If nobody comes up with one I will make one up myself and
> post it to the list (and RMS) for OK.
>
> Cheers,
>
> Mark


--
Etienne M. Gagnon                    http://www.info.uqam.ca/~egagnon/
SableVM:                                       http://www.sablevm.org/
SableCC:                                       http://www.sablecc.org/