[Neurodebian-users] [Nipy-devel] nipy-suiteness

[Gael Varoquaux]

On Sun, Feb 20, 2011 at 11:26:57AM -0500, Fernando Perez wrote:
> On Sun, Feb 20, 2011 at 9:50 AM, Yaroslav Halchenko
> <debian at onerussian.com> wrote:
> > We are just humble servants at your discretion, and this is just a little
> > helper package creating a NiPy "umbrella" for (Neuro)Debian users; thus I
> > only can ping the major kudos back to you for the nice projects ;-)


> Please don't diminish the value of your contributions: over the last
> 10 years, if there's one thing that *consistently* comes back as the
> major hurdle for greater Python adoption in scientific computing, it's
> installation/deployment hassles.  That's why the advent of EPD made
> such a huge difference when it solidified, and your work is on the
> same lines.  It's time consuming, not always fun and critically
> important.  Of course, you couldn't package something if it wasn't
> written in the first place, but the packaging part is not just a
> footnote.  Great job, and much appreciated.

+1 on all that. I would even add that it extend beyond Python. A
roadblock to bringing an scientific computing library (and not turn-key
program) to the masses is distribution. I actually think that it is
harming computational science significantly.

Gaël