[Quantian-general] scipy problems

Dirk Eddelbuettel edd@debian.org
Wed, 3 Mar 2004 19:41:32 -0600

Hi Ulf,

On Thu, Mar 04, 2004 at 02:13:49AM +0100, Ulf Lorenz wrote:
> Hi,
> we wanted to use Quantian for a course in computational physics.
> Although all lecturers have intensively lobbied for Linux, some people
> still need some advocacy. So beside a windows CD for the real windows
> users we wanted to supply the participants with an easy-to-use Linux
> distribution. :)


> However, one small problem. I have just booted quantian another time and
> found out that this problem isn't as, ehm, problematic as I thought on
> my first try. Imagine the following commands in the python prompt:
> >>> import scipy
> >>> import scipy.xplt
> >>> from scipy import *
> >>> from scipy.xplt import *
> >>> x = scipy.arange(10)    #just to create an array
> >>> scipy.xplt.plg(x)       #works
> >>> plg(x)                  #works
> >>> xplt.plg(x)             #doesn't work, some errors with "cannot
>                              import ..."
> Since the homework relies on scipy.xplt excessively, this is not really
> great. Just now I have found out that the first two cases (which I
> hadn't tested before) work, so this problem is not critical as one could
> tell the people to use xplt _this_ way, but it is a bit ... dirty.
> I haven't dug much more apart from the "this shouldn't happen"
> enlightment and will have a deeper look tomorrow, but I wanted to
> send this mail as soon as possible just to inform you. Maybe we will
> have some further requests later on. :)

I can't really help on this as I don't actively use SciPy, but I will CC
this to Jose who may have an idea.

Jose:  Thanks for repository. I use it to augment the Quantian live cdrom I
am providing (http://dirk.eddelbuettel.com/quantian) which contain knoppix
plus openMosix extensions plus a host of scientifc software (some of which I
maintain for Debian), plus a few from you.  As you can see in Ulf's email,
he has questions.  Could you comment on this? Are there workarounds?

> Ulf
> BTW: Could you CC the mail to me? I am not inscribed to the list yet
> (maybe later).


Regards, Dirk

The relationship between the computed price and reality is as yet unknown.  
                                             -- From the pac(8) manual page