Speeding up Python

[Ionel Simionescu]

| >If you are bound to the use of quite unusual data structures, which make
| >difficult vectorization, then you may need to use Python for prototyping
| >only. However, no widely used math package or language can help you
better
| >in such a case.
| .
| Now *that* surprises me.  Maybe I've misunderstood
| what you've written here; do you truly prefer Python
| to Mathematica and Matlab for, say, magnetohydrody-
| namic or relativistic gravitational field solutions?
| Now this has me thinking ... we've certainly compared
| "scripting languages" often enough.  It's past time
| to make explicit how well Mathematica and NumPy do
| common engineering calculations.  The first thing I
| realize in thinking about this is that I'd write
| quite differently in the two; object orientation
| really does make a difference.
| --


When writing I had in mind original problems, that is ones for which there
is no crystalized formalism/implementation/library support. Thus, I wanted
to emphasize the suitability of Python for design.

For traditional problems I would strongly recommend the use of the language
which offer the best libraries for your problem, overseeing (at least in
part) the pure language issues. What matters more is the problem to solve.

I did not mean to compare Python to Mathematica. Mathematica is powerful and
nice as a language and has a tremendous library suppport. However, I dislike
highly the mess where you are thrown in when it comes to debugging, and the
fact you are completely bound to their world - no natural way out to other
languages, commercial products, etc.

Indeed, Python helps for a 'different' frame set. However, I guess the
reason for this is not as much in the language itself, as it is in the
freedom you are given. Making programming almost natural speaking and
easilly capturing complex language constructs, Pythons frees one from the
hassle of observing lots of rules, and thus allows him to become more
creative. But this is just a matter of releasing a potential. Positive
effects show up if the potential exists. One can write ugly code in Python
equally easy. And I am afraid most will do so. Just wait for Python to go
high with the 'public' and see then. Actually I am afraid this will happen
real soon.

ionel