I have successfully (I think) coerced my students into using

Scipy/Numpy for signal processing and dynamic system modeling. They

are mechanical engineering coming from a Matlab background. In order

to make using Python easy and have it feel like Matlab, I teach them

to put

from scipy import *

from pylab import *

at the beginning of every script. One problem with this is that the

zeros and ones functions of pylab default to integers while those from

scipy default to floating point numbers. I have had several students

this week beating there heads against problems that turned out to be

trying to put floating point values in an array that was created using

pylab.zeros.

Can this be changed? Is there a better approach I should take in

getting my students started using scipy and pylab together?

Thanks,

Ryan

Yes, this was left in initially for backward compatibility but I think

we should strive for maximal numpy compatibility going forward. In

svn, we now have pyplot ehich is like pylab but minus the numpy stuff.

Thus if you are using svn (or future releases) you can encourage your

students to

from scipy import *

from matplotlib.pyplot import *

## ···

On 10/11/07, Ryan Krauss <ryanlists@...287...> wrote:

Can this be changed? Is there a better approach I should take in

getting my students started using scipy and pylab together?

Can this be changed? Is there a better approach I should take in

getting my students started using scipy and pylab together?

Teaching them what is a module and a namespace is a good think, why don’t telling them to do :

import pylab as pl

import numpy as npy

for instance so that they have a feeling it’s almost like MAtlab, but better structured ?

Matthieu