For Approval: Open Source Software Alliance License

[Brian Behlendorf]

On Sat, 27 Sep 2003, Lawrence E. Rosen wrote:
> I'm sorry that I'm coming in late to this conversation but I've been busy.
>
> I'm concerned about the following section of the proposed license:
>
> 4. Redistributions of source code must not be used in conjunction
>    with any software license that requires disclosure of source
>    code (ex: the GNU Public License, hereafter known as the GPL).
>
> Licenses seem sometimes to be used as weapons rather than to foster freely
> reusable code.  In this case, the author has made clear, he wants to allow
> his software to be used with proprietary derivative works but not with the
> GPL.  It is, I guess, the anti-GPL license.
>
> In that sense, I think, it violates the OSD.

Which term?  The "discrimination" clauses, #4 and #5, talk about "persons
or groups", and "fields of endeavor".  Defining what either of those mean
has always been troublesome.  An easy test is nationality (can't deny
Polish citizens, for example) or career (can't deny the military).  It's a
big stretch to try and claim that "people who use the GPL" is a "group" or
"field of endeavor" - because if we did, every term of every license could
be challenged as OSI-incompatible.

I suppose we could point to #9, and claim that the word "use" in the above
section is too vague and could apply to mere aggregation on a single
distribution medium.  Or that "use" in the context of using OSSAL software
on a GNU-licensed operating system is also too vague.

But as far as I can tell, the proposed license is "anti-GPL" in the same
way that the GPL itself is anti-everything-that-isn't-a-subset-of-the-GPL.
We may not think it's a very good license, most of us would probably lobby
against its use, but I can't see a term of the OSD that it violates (at
least the section above, I've not read the full OSSAL license).

Whether Sean realizes it or not, I think Ernie Prabhakar captured Sean's
primary concern most closely: that a GPL-only fork would arise and capture
more development momentum than a BSD-licensed original.  I can't recall
that ever happening, but I do hear similar sentiments from Apache members
from time to time.  OpenOffice has had issues with developers fixing bugs
or adding new features but only offering those patches under the GPL,
rather than granting them to Sun to dual-license under GPL and SISSL.

> On 6/30/2003 Brian Behlendorf asked this list to consider a recent Microsoft
> license provision that read as follows:
>
>   [You agree] [t]hat you are not allowed to combine or distribute the
>   Software with other software that is licensed under terms that seek to
>   require that the Software (or any intellectual property in it) be
>   provided in source code form, licensed to others to allow the creation
>   or distribution of derivative works, or distributed without charge.
>
> Once again, this is a license that says "don't use that license" rather than
> "do use this license."  The former wording seems like discrimination to me
> and the latter like any reciprocal license we've approved since the
> beginning.

I don't recall a consensus being formed as to whether the above violated
any OSD term.  Of course I'm not pushing for that license to be certified,
neither do we see anyone from Microsoft here to push it, but it's still an
interesting test case IMHO.

> Is that a distinction without a difference?  Or should we assert that
> licenses of the form "don't use that license" are contrary to the OSD
> because they discriminate?

I would wonder whether you can reasonably define "use" above in a way
that doesn't de-qualify the GPL.

	Brian
--
license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3