<html><body>
<p>I think the Forum should make it clear that a reference implementation may not carry a GPL license. <br>
<br>
Any GPL like license would greatly restrict the number of people who could evaluate the implementation<br>
<br>
<br>
Dick Treumann - MPI Team <br>
IBM Systems & Technology Group<br>
Dept X2ZA / MS P963 -- 2455 South Road -- Poughkeepsie, NY 12601<br>
Tele (845) 433-7846 Fax (845) 433-8363<br>
<br>
<br>
<tt>mpi-forum-bounces@lists.mpi-forum.org wrote on 04/03/2009 03:17:58 PM:<br>
<br>
> [image removed] </tt><br>
<tt>> <br>
> Re: [Mpi-forum] MPI prototype license</tt><br>
<tt>> <br>
> Jeff Squyres </tt><br>
<tt>> <br>
> to:</tt><br>
<tt>> <br>
> Main MPI Forum mailing list</tt><br>
<tt>> <br>
> 04/03/2009 03:21 PM</tt><br>
<tt>> <br>
> Sent by:</tt><br>
<tt>> <br>
> mpi-forum-bounces@lists.mpi-forum.org</tt><br>
<tt>> <br>
> Please respond to Main MPI Forum mailing list</tt><br>
<tt>> <br>
> On Apr 3, 2009, at 3:04 PM, Supalov, Alexander wrote:<br>
> <br>
> > Many prototypes are being developed currently for MPI-2.2 and MPI-3. <br>
> > Under what license will they be made available to the community? <br>
> > Certain types of licenses, like GPL, may not be suitable for <br>
> > commercial adoption of the prototype code. BSD-like licenses are <br>
> > normally acceptable. I think we should discuss this matter at the <br>
> > Forum this time. You comments and suggestions are welcome.<br>
> <br>
> I guess that raises the meta question -- what is the intent? Is it a <br>
> requirement that the proof-of-concept open source implementations must <br>
> be licensed such that, in practice (not theory) they can be slurped <br>
> into commercial/closed source MPI implementations? Or is it enough <br>
> that some people implemented it, some [other] people reviewed it, and <br>
> the union of those people have said "yes, this shows what we want it <br>
> to show"? (and potentially a 3rd set of people who didn't look at the <br>
> proof-of-concept implementation go re-implement it in their MPI <br>
> implementation)<br>
> <br>
> I thought the intent was to show that the concept was possible and to <br>
> help unearth any "gotchas" in terms of implementation that might not <br>
> be thought of during the design phase.<br>
> <br>
> But to be clear: OMPI, where I have done all my implementation work <br>
> for MPI 2.2 tickets, is BSD (IANAL, and none of this is legal advice <br>
> -- go read our license yourself) so I don't really care too much.<br>
> <br>
> -- <br>
> Jeff Squyres<br>
> Cisco Systems<br>
> <br>
> _______________________________________________<br>
> mpi-forum mailing list<br>
> mpi-forum@lists.mpi-forum.org<br>
> <a href="http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum">http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum</a><br>
</tt></body></html>