[Pkg-exppsy-pymvpa] model selection and other questions
emanuele at relativita.com
Wed Apr 2 22:06:04 UTC 2008
Michael Hanke wrote:
> I completely second Yarik's statement (thanks for being verbose),
> therefore I will jump directly to the last open question.
> Emanuele, if you decide to contribute (some of) your code to PyMVPA, you
> can have commit access at any time. We'd also be very interested in your
> oppinion about PyMVPA's class hierarchy and where your code would fit
> best (or what refactoring is needed to make it fit).
Excellent. I need a bit of time to understand the whole picture
and to become decently proficient. Thans for this kind invitation.
>>> Which is the reason of MIT/X11 license instead of (modified) BSD or GPL?
> This is a surprisingly complicated issue. The original choice of MIT was
> done to be compatible with NumPy/SciPy world. Both actually are fine
> with BSD and MIT:
> At some point shogun appeared on the radar and caused licensing trouble.
> This trouble is the major reason why shogun interface is not yet part of
> the PyMVPA main branch.
> However, currently I'm not convinced that this is a real problem as
> shogun provides a plain python interface with sufficient functionality,
> so we do not have to link against any shogun lib. We still need to
> clarify this however.
> But as Yarik said, all copyright holders (which are few) are in close
> contact and could potentially change the license in any direction. We
> already had one request to change it to GPL. But currently, IMHO there
> is no need to switch, besides philosophical reasons (which I nevertheless
> consider valid) and I personaly prefer to stick to the current license
> unless some situation prevents us from doing what we want to do ;-)
> Hope that explains.
Good explanation. I'm still surprised by the choice of X11/MIT over
BSD, but it is more a curiosity. I have the highest opinion of GPL
but I somewhat understand why NumPy/SciPy developers did not
More information about the Pkg-exppsy-pymvpa