[SciPy-user] OS-X Universal binary?

[Christopher Barker]

Sebastian Haase wrote:
> This of course reminds of a fundamental discussion from long ago -- as
> I remember there were plans to split scipy into sub-packages that
> would possibly be able to live by themselves and make the whole thing
> easier to install.

yup.

> Anyway, even if this plan is not to be followed through, I would like to know:
> * What are "easily separable packages" ? I would think ndimages might
> be an examples, or ?
> * How much really depends on fortran ?  This is also interesting to
> get a "quick" win-64 built.

I think that's the trick -- the fortran depedencies are not alwyas 
clearly separable...

In the past, I've found it a bit tricky to have a package that could be 
installed separately, but when installed, it would be just part of a 
bigger package -- that's not to say it can't be done.

>   [ Chris, which imports did you comment out and where ? ]


I don't have that machine handy, but it was in the scipy.interpolate 
package. In it's __init__.py, it imports a bunch of stuff, some of which 
is (or relies on ) fortran. I think it was mostly fitpack.

The trick is that is is nice to have the fancier interpolaters right 
there with the rest.

-Chris