[SciPy-user] Matrix docs?

[Skip Montanaro]

    >> but it seems odd to me that this isn't the form shown in the
    >> tutorial.
    >> 
    >> 
    Travis> Why is this odd?  The list of lists approach is extensively
    Travis> documented in the numpy manuals, but this "string approach"
    Travis> which will be more familiar to matlab users is useful and so is
    Travis> documented in the tutorial.

Yes, but this is in the Scipy Tutorial, which as far as I've been able to
determine makes an assumption that someone new to SciPy will be experienced
with Numeric Python.  That's certainly not the case for me (an experienced
Python programmer, but not a scientific programmer), nor will it be the case
for our researchers.  Some have used Python.  None have used SciPy or
Numeric before.

It would be really helpful if the first statement of the SciPy Tutorial had
a footnote referring the reader to Numeric Python documentation.  (Is the
NumTut package a Numeric tutorial?  Is it available by default through
scipy?)

    >> I eventually stumbled on the NumPy docs.  
    >> 
    Travis> The NumPy docs are a great source since scipy just adds features
    Travis> to NumPy.  I would suggest always starting there.

Sure, now I realize that.  But as a rank beginner I started with the SciPy
tutorial which didn't say, "Stop! We assume you're already familiar with
Numeric Python.  Better check out blah blah blah..."

    Travis> The Enthought people have a .chm file which contains a great
    Travis> deal of information and livedocs is helpful provided there are
    Travis> docstrings (which is mostly true throughout scipy --- but not
    Travis> throughout Numeric as you've noticed, the mat documentation
    Travis> comes from Matrix.Matrix which is sadly lacking).

All the more reason to strongly direct newcomers to the Numeric docs.

Thanks for the assistance,

Skip