Should i pick up Numeric, Numarray or SciPy.core?

[jelle]

#No rant intended

I'm not at all confused wether I should learn an one of the advanced
array modules, I'm slightly confused over which I should pick up. I'm
impressed with the efforts of SciPy and Scientific, but since I'm
fairly new to programming & OO, choosing Numarray over Numeric hasnt
been that satisfactory at all, which is not a problem of the quality of
Numarray, but the fact that a lot of modules still heavily realy on
Numeric. Which, with my fragile knowledge in programming is rather
confusing sometimes. Since its seems its far from pragmatic to opt
solely for either one of them. So Scipy.core looks like the ticket for
me here, doesnt it? Before putting an effort into it, I'd like to know
whether this would make sense, certainly since scipy.core is in its
early stages, and documentation isnt available yet (for free that is...
its not clear to me why I should pay for software in its early stages,
when it extents on efforts freely available)

Curious for your opinions, I definitely think the Scipy.core effort is
terrific, but right now I'm wondering whether its a pragmatic choice.

-Jelle

 

[Robert Kern]

#No rant intended

I'm not at all confused wether I should learn an one of the advanced
array modules, I'm slightly confused over which I should pick up. I'm
impressed with the efforts of SciPy and Scientific, but since I'm
fairly new to programming & OO, choosing Numarray over Numeric hasnt
been that satisfactory at all, which is not a problem of the quality of
Numarray, but the fact that a lot of modules still heavily realy on
Numeric. Which, with my fragile knowledge in programming is rather
confusing sometimes. Since its seems its far from pragmatic to opt
solely for either one of them. So Scipy.core looks like the ticket for
me here, doesnt it? Before putting an effort into it, I'd like to know
whether this would make sense, certainly since scipy.core is in its
early stages, and documentation isnt available yet (for free that is...
its not clear to me why I should pay for software in its early stages,
when it extents on efforts freely available)

The software itself is absolutely free. The complete documentation for
scipy_core specifically is available for a fee, but you will get free
updates as the documentation is expanded and as soon as a certain number
of copies have been purchased or a certain amount of time has passed,
the documentation will be made available free of charge. The book does
not extend "on efforts freely available," either; it's entirely new.

However, not much has changed between Numeric and scipy_core, so most of
the Numeric documentation still applies. There is fairly good docstring
coverage in any case.

Curious for your opinions, I definitely think the Scipy.core effort is
terrific, but right now I'm wondering whether its a pragmatic choice.

scipy_core is the future. Most of the complete scipy package has been
ported over at the moment.

However, if you need to use other code that depends on Numeric (like
Konrad Hinsen's Scientific), then you can use Numeric for the time being
and port your code to scipy_core using the provided conversion script.
The API on the Python side hasn't changed too dramatically.

--
Robert Kern
(e-mail address removed)

"In the fields of hell where the grass grows high
Are the graves of dreams allowed to die."
-- Richard Harter