[Pkg-octave-devel] Bug#435214: Bug#435214: Bug#435214: Clarification of licensing terms of semidef-oct

[Rafael Laboissiere]

* Stephen Boyd <boyd at stanford.edu> [2007-07-31 16:41]:

> i (and many others too) do most of my day to day convex optimization 
> work using cvx, which has a GNU license, but unfortunately, runs on top 
> of matlab.  our plan (hope? goal?) however is to develop a full GNU 
> system for modeling and solving convex optimization problems.  lieven's 
> CVXOPT is a good start along these lines, but i am not sure how easy or 
> hard it is to tie python/C into octave.  if you are interested in 
> optimization software for octave, i'd be happy to chat; maybe we can 
> help (beyond changing the license on ancient, but working, software to GNU).

Thanks for your prompt reply.  I appreciate your efforts for clarifying the
license terms of semidef_prog, together with .

I suspected that semidef_prog was already an old albeit working software.  I
do not use optimization software very frequently and my interest is to
insure that the octave-sp package (derived from semidef_prog) will not
disappear from Debian.  Note that the number of users who already installed
the package is non-negligible [1].

As regards your proposal to get your more recent software ported to Octave,
I think it is a great idea.  The 2.9.* series of Octave has a improved
Matlab compatibility level, so the chances your packages will run on Octave
are increased.  Moreover there is already a group of developers of the
Octave-Forge project [2] working on optimization. They might be interested
in including your new software in their package [3].

At any rate, the more urgent thing to do is to clarify the semidef_prog
license terms, such that the package can be kept in Debian.

[1] http://qa.debian.org/popcon.php?package=semidef-oct
[2] http://octave.sourceforge.net/
[3] http://octave.sourceforge.net/optim/index.html

Regards,

-- 
Rafael