[SciPy-dev] Inclusion of cython code in scipy

Matthew Brett matthew.brett@gmail....
Wed Apr 23 02:12:06 CDT 2008


Hi,

>  If using Cython isn't acceptable to the SciPy/NumPy community at large
>  (i.e., if I'm going to work on something for 2-3 months that won't
>  ever be committed), then it would be nice to know that up front.  I
>  can either choose to use other tools, or do some work to help make
>  using Cython less of a pain.  Either way is fine with me, of course.

It seems to me likely, especially from what I hear about cython in
SAGE, that cython is going to be a very important tool for scientific
python, and could well be absolutely central.  Python makes code
readable, and maintainable, and yet, for scientific applications, we
often need the speed of C or its libraries.  There is a much smaller
number of Numpy / Scipy developers who feel comfortable working on the
Numpy C code, and that means fewer eyes reviewing and maintaining
large parts of Numpy. Cython has the potential to make writing C with
Numpy very nearly as easy as writing python.  If that happens, it
could make it an order of magnitude easier to develop with Numpy, and
that in turn could really change the speed and quality of development.
 So, I ask, on bended knee, please, go for it.  Many of us are very
excited to see how your project goes. .

Cheers,

Matthew


More information about the Scipy-dev mailing list