New license - please comment

[Mark Rafn]

On Tue, 10 Jun 2003, Christophe Dupre wrote:

> The problem (from our perspective) with your proposition is that if the
> 3rd party is too lazy to submit a patch, is run over by a bus, or is
> unavailable, even if we get the code from a 4th party that got it from the
> 3rd party, we're no longer able to re-license the code to commercial
> entities.

Indeed.  Also, you're not able to re-license other people's work in the
case where the 3rd party doesn't want you to do so, and she instead just
forks the project without granting you special rights to it.

That's among the strengths of open-source - people choose how and if they
will contribute.

> The type of libraries we're planning on releasing are not general public
> code (as zlib, or gtk, or other could be). They're very specific libraries
> for parallel computing on teraflop-sized machines.

That's what they are now.  Among the other strengths of open-source 
software is the fact that it can be extended far beyond the expectations 
of the original author.

> They are the result of several years of research, and through the years
> we've found that our biggest difficulty is to get patches in from people
> who've extended them. Most of the work is done by grad students, who
> couldn't care less about what happens to their work after they graduate.

This is unfortunate, but it is their right.  I don't know if a more 
heartfelt plea would help, but I do know that the license hammer is not 
the appropriate way to address it.

> Now, what is inherently bad about clause 2(d) ?

An argument was made on debian-legal that this amounts to a fee for the 
right to make derived works.  "You must pay us $1 to make a derived work" 
is not that different from "You must give us any value you add via a 
derived work".  I wonder if OSD#7 (the demand to relicense derivative 
works is, in fact, a seperate license on other software) applies as well.

> We're providing the libraries freely (money). And you are free to use
> them in whatever programs you want. The only thing we ask is if you
> modify them, we co-own the copyright. You still keep it, which gives you
> a garantee that there's always going to be a free version out there
> (i.e. we can't just take it back).

I'm not saying this is a horrible evil thing to want, only that it falls 
(in my mind) on the proprietary side of the fine line between open source 
and proprietary.  As proprietary licenses go, it's very liberal.

> Giving us co-ownership of the modifications doesn't take anything from
> you, does it ?

Sure it does.  It prevents me from knowing that my derived work is fully 
free software.  It prevents me negotiating with you for a cut of the take 
if you want to sell my work. 

If you just said "This work and all derived works are co-owned by every 
legal entity", would it take anything from you?

> It doesn't restrain your freedom. It's no different from the GPL that
> makes you give out your code if you link your proprietary (closed) code
> with a GPL library (not LGPL).

Aside from the fact that that would be an unlikely remedy for most GPL
infringements, the GPL does not demand that I give someone the right to
relicense my work (or derived works of my work) in such a way that I
cannot use it.
--
Mark Rafn    dagon at dagon.net    <http://www.dagon.net/>  
--
license-discuss archive is at http://crynwr.com/cgi-bin/ezmlm-cgi?3