[SciPy-user] download and install python with all modules

Robert Kern robert.kern@gmail....
Tue Jul 24 14:15:51 CDT 2007


Stephen McInerney wrote:
> a) A very good zero-hassle batteries-included Python distribution for 
> Windows
> with scientific and visualization libraries is Enthought Python:
> http://code.enthought.com/enthon/
> Current release is Python 2.4.3 ; (Python 2.5.x is behind but supposedly due 
> "this August").
> Enthought Python 2.4.3 has NumPy 0.9.9.2706 + SciPy 0.5.0.2033 + IPython 
> 0.7.2 + Matplotlib 0.87.3.2478 + wx +much more ...
> Typically Enthought bundles lag Python releases by ~1 yr.
> You might contact enthought-dev@mail.enthought.com yourselves to ask for 
> 2.5.x :)

We're not updating the monolithic distribution anymore. Instead we are building
binaries for each package separately and using Python eggs as a distribution
mechanism.

  http://code.enthought.com/enstaller/

> b) Compiling 15+ individual packages on Windows and reconciling multiple 
> version conflicts
> is a huge pain and a barrier to adoption, in particular if we need the retro 
> version
> Visual Express C++ for 2003. It's just not going to fly.
> 
> To add a question I have always wanted to know, are there any firm plans to 
> move
> NumPy + SciPy into Python?

None. There is a PEP for putting the C array interface into Python for efficient
communication of multidimensional array data between C extensions. In a few
years, we might be able to contribute a slimmed down array type and ufuncs for
use at the Python level, but that seems less likely. scipy will never be a part
of Python.

> It really needs an arbitrary-precision numercial 
> type (no, 'Decimal' is not good enough, too slow and basic methods like pow 
> and math.* don't work on it )

We don't have an arbitrary-precision numerical type either.

-- 
Robert Kern

"I have come to believe that the whole world is an enigma, a harmless enigma
 that is made terrible by our own mad attempt to interpret it as though it had
 an underlying truth."
  -- Umberto Eco


More information about the SciPy-user mailing list