[PYTHON MATRIX-SIG] Data analysis package question

Perry A. Stoll pas@lems.brown.edu
Thu, 31 Oct 1996 15:33:21 -0500 (EST)



On Thu, 31 Oct 1996, Geoffrey Furnish wrote:

> Jim Hugunin writes:
>  > > I have been exploring the possibilities of data analysis using
>  > > Yorick and Python.
> 
> It is worth pointing out that PLplot fits into the above picture as
> well.  I understand that my forbears have been in the habit of

I'm going to relate my opinion/experience concerning plplot as it relates
to NumPy. 

PLplot is a bear to compile. I've had a fair amount of experience working 
with large software projects, and getting PLplot to the point where I 
could use it, in or out of python, took a significant amount of work. Way 
too much, in my opinion. And it even uses gnu configure.

PLplot needs to be better integrated into Python. Namely, it should _not_
simply exit when it can not allocate a desired color. This behaviour has
forced me to stop using it. I'd be doing some analysis in a Python
interpreter, decide I'd like to visualize the result and import plplot.
I try to open a plot and it quits with a message along the lines of 
"couldn't allocate color FOOBAR. Quitting..." losing all the data I'd 
spent time working to get. 

There was very little documentation or support information. Such as which 
version of plplot (4.99alpha.k.1.4.5.6  or 4.99.gamma.j.3.4.2.1) the 
given plplot module worked with.

Ok, you say, we'll gladly accept patches to fix it. I did submit some
small patches to the author of the PLplot module, but I don't have the 
time or inclination to make any more. 

My general impression was that PLplot was overkill for my simple
visualization needs - plots, contour maps and surfaces. I guess I'm just
saying to those considering using PLplot from within Python - it's not an 
easy road to travel.


> My own personal opinion on what is most needed in the Python universe
> in order to support such effots by myself and others, are:
> 
> 1) Thorough documentation for the NumPy C interface.
> 2) A C++ wrapper for NumPy data storage (I have a start on this, which
>    I should be able to release soemtime).
> 3) A real plan for doing serious "megawidgets" in Python/Tk.

Sounds good, especially 3. I think light weight Py/Tk widgets would be 
the way to go.

-Perry
 <pas@lems.brown.edu>

=================
MATRIX-SIG  - SIG on Matrix Math for Python

send messages to: matrix-sig@python.org
administrivia to: matrix-sig-request@python.org
=================