Hi Paul,
where did you get your ideas from?
Did you read and interpret the GPL yourself or did you read the incorrect
GPL FAQ from the FSF?
Paul Jakma <[EMAIL PROTECTED]> wrote:
> On Wed, 2 Aug 2006, Joerg Schilling wrote:
>
> > The GPPL however allows to use code of _any_ license from a GPL project.
>
> No it does not, for the GPL demands that derived works be licenseable
> under the GPL without additional restrictions (see section 6). Code
> under the CDDL can not be taken and licensed under the GPLv2, for the
> CDDL contains restrictions (e.g. the patent clauses) which the GPLv2
> does not.
This is not true for several reasons (in case you do not mix
code from both licenses in a single file):
- Both licenses are source licenses and to not claim anything
about the "license" of a binary. They ony require that the
source needs to be available in case you distribute a bunary.
- The CDDL is file based and does not put any restrictions on the
GPLd code.
- If the GPL would try to put any restriction on the CDDLd (or any
other code), then at least this part of the GPL would be illegal.
Note that in addition, a German (and most likely any European) has the right
on the "wissenschaftliches Kleinzitat". This allows to use small portions of
other people's work without permission, in case that you put a note on
the original Author in the code. This even allows to mix code in a single
file (in case that the parts taken from other authors does not form a
significant amount on the "creation" of the whole).
> I.e. there is no way to create a work derived of both GPLv2 and CDDL
> code without failing to violate /one/ of the licences. Ergo there is
> no way to combine CDDL and GPLv2 code.
I get the impression that you have a far too wide view from a "derived work".
The fact that the GPLv2 uses the term "derived work" makes it hard to
discuss it.
If you like to find out what is possible, you should start with cases where
clearly no "derived work" is created. Note that the GPL does not contain
the term "linking" at all and for this reason, simple linking does not create
a "derived work".
Once you did understand what the GPL alows in such cases, you have the base
for thinking about more difficult cases.
> That's a pretty widely recognised incompatibility really.
This is the well known FUD from certain people...
> For cases where code is combined but no derived work is created, the
> whole licence question is, quite obviously, moot and irrelevant.
> However, see below about derived works.
And the term "derived work" needs to be interpreted the right way.
This only can be done after you did understand which cases clearly do not
create a "derived work".
> > Mkisofs is e.g. a GPLd project. It uses libscg and libschily which
> > are both under CDDL. As this fact does not make libscg/libschily
> > derived works from mkisofs, there is no problem with this
> > combination.
>
> Well, what would and what would not constitute a derived work is
> something one really would need to seek legal advice on.
The claim from the FSF that you only need to put one single line of GPLd
code into your work in order to make it a "derived work" is definitely
wrong as this is an illegal (against law) claim.
> My vague understanding is that there are no easy rules of thumb
> laymen can follow other than to exercise caution when combining code
> and to assume the resulting code /does/ derive, unless one hears
> otherwise from professional counsel.
This is why I would like to see a _useful_ GPLv2 FAQ from the FSF
with a higher priority than a GPLv3.
> In your example above, there is also a question of whether anyone
> would care. I'm guessing that either:
>
> - the author is the same for both the libraries and mkisofs
> or
> - that the mkisofs authors chose to use CDDL libraries
>
> Which is highly unlikely to ever lead to any sustained legal
> problems, for several reasons.
Well,there are some trolls from a well known site that claims to be
a torch-bearer of "free software" that "care" about this and claim that
they cannot legally redistribute this project, which is obviously FUD.
> > The problem is that the FUD is that you cannot combine CDDL & GPL code...
>
> It is not FUD though I'm afraid.
If you believe that it is not FUD, then please tell me why the mentioned
combination with mkisofs would not be OK. Note that I am not the 100% author
of mkisofs.
> > Why do you write this?
> > Do you believe that we need GPL code for OpenSolaris?
>
> If the code is useful, why not? The GPL and CDDL aren't very far
> removed. And hopefully CDDL and GPL will become compatible with the
> GPLv3, which should finish off a lot of these questions.
I did not yet read the 2nd version of the GPLv3 draft, but in the first
version, I could not find any explicit sentence that mentions mixing with
code from other licenses. Well, Eben Moglen claims that the GPLv3 allows it,
but then I am sure that the GPLv2 already does in many cases.
> > The other direction of code flow is already permitted
> > and using GPLd code from OpenSolaris may also possible in case that
> > OpenSolaris will not become a "derived work" from the GPLd
> > code.
>
> Which is very difficult to determine generally, see points above. It
> all becomes moot if/when GPLv3 allows compatibility.
In any case, the CDDL is a free license that allows a combination
with other licenses (even BSDl). I am in hope that the GPLv3 will finally
be explicit enough with permitting code mix with other free licenses.
Jörg
--
EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin
[EMAIL PROTECTED] (uni)
[EMAIL PROTECTED] (work) Blog: http://schily.blogspot.com/
URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily
_______________________________________________
opensolaris-code mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/opensolaris-code
