Scientific Libraries in Python

[Fernando Pérez]

> 
> Would it be possible for one to absorb the other so we have a single
> standard scientific library for Python?

I think the following I saw today on sci.math.num-analysis is worth 
mentioning in the context of this discussion:

------------------copy

The first major release of the GNU Scientific Library is now available
for general use.

The GNU Scientific Library (GSL) is a free numerical library for C
programmers.  The routines have been written from scratch by the GSL
team in ANSI C.  They present a modern API, allowing wrappers to be
written for very high level languages.  The library includes a
450-page reference manual and is distributed under the GNU General
Public License.

The source code can be downloaded from ftp.gnu.org in the directory
/gnu/gsl/ or from any mirror site (see http://www.gnu.org/order/ftp.html)

The project home page is at http://sources.redhat.com/gsl/

------------------/copy

To avoid unnecessary duplication, it might be worth considering a 
SciPy/Scientific/whatever system to be based on something like the GSL. This 
has some important advantages, IMHO:

- Because the base code is not python-specific, many more people will 
contribute to it/debug/enhance it.

- The GSL guys specifically mention a wrapper-friendly design of their 
project. So maintaining the python interface shouldn't be all that hard 
(famous last words...)

I'm not saying that a Science Python environment should be *only* a wrapper 
around GSL, but that the GSL should be one core component of such a project. 
Other things can then be added to from the full environment (like 
visualization and the ultra-cool scipy.compiler system).


Cheers,

f.