[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [pygame] 3D SHMUP



Thanks for fielding this Miriam. Your account is pretty much how I understand it. I have found a history of the various incarnations of the array module here:

http://www.scipy.org/History_of_SciPy

Miriam English wrote:
Ian Mallett wrote:
On 9/16/07, *Lenard Lindstrom* <len-l@xxxxxxxxx <mailto:len-l@xxxxxxxxx>> wrote:

    NumPy is not the same as Numeric. As long as you are using a 32 bit
    machine you can compile Numeric for Python 2.5.

I thought so. I've heard things about development on Numeric had stopped, and was therefore dead?

The whole Numeric/numarray/NumPy screwup makes my brain explode. I keep getting it all mixed up and I expect I'm not the only one.

I think this is how it goes (I might be wrong on some of this):

They are all, as far as I can see almost the same thing, with slight changes, the main result of which, is to break stuff.

#  numarray has been pretty-much dropped.

Like Numeric, it is being replaced with NumPy.

# Numeric was the original, which was upgraded to numarray, but then the old Numeric was re-upgraded to be a better version and released as Numeric again. There are moves to have this dropped.

Numeric did not meet the requirements for inclusion in the Python Standard Library. Numarray was written as a replacement. Though fast with large arrays it proved slower than Numeric for small arrays.

# NumPy was (I think) yet another rewrite of Numeric. It upset a lot of people by having some of the documentation released non-free, but there is considerable pressure to make this the main version. Nowadays there is quite a bit of free documentation for NumPy available.
NumPy is an updated Numeric, like Python 3.0 is to Python 2.x. It breaks things but is not a complete rewrite. It incorporates features from numarray as well. The name NumPy was also associated with Numeric in the past. How is that for confusing?

  http://www.scipy.org/Documentation
  http://www.scipy.org/doc/numpy_api_docs/numpy.html
  http://docs.neuroinf.de/api/numpy/
  http://www.scipy.org/Numpy_Example_List_With_Doc
  http://projects.scipy.org/scipy/numpy/wiki/

The main problem is that even if NumPy is used forever more it doesn't help all the old programs out there that used numeric. This problem will become less annoying given the passage of time. My solution has been to install both Numeric and NumPy. For those using Python 2.5 René Dudfield has made a precompiled version available:
http://rene.f0o.com/~rene/stuff/Numeric-24.2.win32-py2.5.exe

For all else go to the Sourceforge pages:
http://sourceforge.net/project/showfiles.php?group_id=1369

Hope this helps. If I got any of this wrong can someone correct me please? I'm still pretty confused about the whole thing, and part of my motivation for writing this was to help straighten it out in my own mind.



--
Lenard Lindstrom
<len-l@xxxxxxxxx>