Alan G Isaac
aisaac at american.edu
Fri Jul 15 07:05:47 CDT 2005
On Thu, 14 Jul 2005, Rich Shepard apparently wrote:
> I'm new to python and I'm trying to find my way in the wilderness of
> various packages as well as old versus new documentation. Right now I'd like
> to understand the differences between SciPy and ScientificPython.
> I suspect that my immediate needs for matrix manipulations -- particularly
> the calculation of the principal eigenvector -- may well be met by NumPy or
> Numeric. That's another pair whose differences I don't understand.
> Anyway, clarification of the SciPy/ScientificPython situation would be very
> much appreciated.
Response from a relatively new user:
Unfortunately, the current situation is confusing.
This is how I understand it.
SciPy currently sits above Numeric. It has extensive
facilities for scientific computation.
Numarray was evolved to address some perceived inadequacies
in Numeric---numarray has some technical advantages, but has
a more "under development" flavor to it, while Numeric is
practically static. This is discussed in part on the SciPy
page. The two packages are currently being unified, and
preliminary work on this unification is available.
Thus the Numeric/numarray distinction should go away "soon",
but for the moment it is complicated by a 3rd option.
ScientificPython is something else entirely. It is
a collection of useful objects for scientific computation.
You can use these in any of the above environments. There
is some overlap with functions provided by SciPy.
One last things as you come on board: I recommend Matplotlib
as your graphics package in any of these environments.
More information about the SciPy-user