On Tuesday, May 20, 2014 5:51:27 AM UTC-7, Frank Millman wrote:

> I used it to install IPython, with the following results.
> First I ran 'pip install ipython', which worked.
> Then I read the IPython docs, which gave the following command to install 
> Notebook -
>     'pip install ipython[notebook]'

the reason for this is that the notebook has additional dependencies beyond 
what "plain" ipython requires. adding the [notebook] is supposed to bring those 
in, but apparetnly did not.

So this is either a bug, or maybe the problem is that you had already installed 
ipython, so when you did :
 'pip install ipython[notebook]'

pip looked and saw that you already had it, so did nothing -- what did it 
report? In this caes:

 'pip install -U ipython[notebook]'

might have worked: -U means upgrade even if I already have it.

But I"m glad the rest of the pip installs worked, that's pretty new actually!

Note that the "scipy stack" and particularly extra add-on stuff like 3-d 
visualization, etc, is very hard to build and install -- so a number of folks 
rely on "distributions" that give you all that in one pile:

Enthought Canopy 
Continuum Anaconda

are two such options.

Very odd that pip being hard to figure out might send someone to MS C#:

Is there really a point and clic, out of the box high level interactive 
computational environment that MS delivers for C#  ;-)

Not that we don't need the installation of packages made even easier, but 



Reply via email to