[SciPy-User] Central File Exchange for SciPy

josef.pktd@gmai... josef.pktd@gmai...
Sat Oct 30 13:18:40 CDT 2010


On Sat, Oct 30, 2010 at 1:42 PM, Gael Varoquaux
<gael.varoquaux@normalesup.org> wrote:
> On Sat, Oct 30, 2010 at 07:00:15PM +0200, Almar Klein wrote:
>>    Woaw, I like your enthusiasm! However, let's first establish whether we
>>    should discard Pypi or if we can maybe make it suitable for our needs with
>>    a few changes (assuming that the rest of the Python community lets us make
>>    these changes).
>
> Yes, I am impressed by the positive attitude that this thread is taking.
> Congratulations for that, and for offering time and energy.
>
> I think the idea in general is an excellent one. In addition to what has
> already been said, here are a few remarks, in random order:
>
>    * Don't disregard PyPI for well-maintained packages: we need the
>      non-scientific Python community, let us not break appart from them.
>      Besides, scientific users also need packages to read XML, talk over
>      the network... On the other hand, for simply cookbook-like stuff,
>      I believe that there might be some value in having a scipy-specific
>    * repository.
>
>    * Searching is important, but can be made easily using a Google
>      custom search. For example, I have been impressed by what Fabian
>      did for the scikit-learn's website,
>      http://scikit-learn.sourceforge.net/, on the top right of the
>      webpage. Try it out, it is cooler than it seems.
>
>    * License: let's at least force users to choose an OSI-compatible
>      license. I would try to push them using the BSD, as in my
>      experience many people choose GPL by default, but I would not
>      enforce this choice.

I would strongly recommend (to users) that all shorter code, snippets
and recipes, are BSD by default or made explicit by the user, and that
the license is very easy to see on the web page.
Given that we are writing BSD code and to avoid any conflicts, I
essentially ignore all non-BSD code, for example on the matlab file
exchange.

Josef

>
>    * Self hosting, bandwidth... this should not stop anybody from start
>      wild ideas. http://docs.scipy.org started as a crazy idea hosted on
>      my girlfriend's server (and coded by Stefan, Pauli and herself :P).
>      It migrated to Enthought-hosted servers when it became more
>      'production-ready'.
>
> By the way, speaking of migration of service, now that ask.scipy.org is
> in production, we should add a link to it on the sidebar of the scipy.org
> website, the docs frontpage, and the planet. I can do it for the planet,
> so if someone does it for the scipy.org wiki, I'll just copy the design.
>
> In general, I think that it is important that all these websites be
> linked together.
>
> As Eleftherios would say 'Go, go team'
>
> Gaël
> _______________________________________________
> SciPy-User mailing list
> SciPy-User@scipy.org
> http://mail.scipy.org/mailman/listinfo/scipy-user
>


More information about the SciPy-User mailing list