[SciPy-user] Numerical Achievements in SciPy

[Rick Ratzel]

>    Date: Tue, 01 May 2007 15:45:17 -0500
>    From: Robert Kern <robert.kern at gmail.com>
> 
>    Karl Young wrote:
>    > It seems to me like the CRAN repository for R based statistics packages 
>    > (http://cran.r-project.org/) provides a pretty good model for something 
>    > like this, though maybe there's something I'm missing re. SciPy that 
>    > makes CRAN a bad model. Having CRAN seems to moderate the size of the 
>    > basic R package (and install) which is convenient, and things that get 
>    > heavily tested and used often end up eventually making it in to the core 
>    > R package.
> 
>    We essentially have the CRAN model using the Python Package Index.
> 
>    1. Write your code.
>    2. Package it using distutils.
>    3. Submit the package to the PyPI http://www.python.org/pypi/. It will host your
>    tarballs and eggs and wininst installers, so you don't have to scrounge for
>    hosting.
>    4. Write up a web page about it if the description text that PyPI allows isn't
>    enough for your code. Feel free to use the www.scipy.org wiki if you like.
>    5. Announce your package here and other places of interest.
> 

   I hope I'm not stating the obvious, but another advantage of PyPI is that all
easy_install users will have access by default.


-- 
Rick Ratzel - Enthought, Inc.
515 Congress Avenue, Suite 2100 - Austin, Texas 78701
512-536-1057 x229 - Fax: 512-536-1059
http://www.enthought.com