When Python's symbolic computation abilities are even at a tenth of Mathematica's (especially for discrete functions like we love in computer science), I'd be more than happy to switch over. Perhaps someone could integrate Python with Sage or Octave.

holy heck, I think I'm going to start doing things in sage. Thanks!!!

not even. I'm working on a mathematica notebook that (mathematically, not physically) simulates the action of an enzyme... It generates a 120x120 matrix of kinetic parameters, symbolically solves to make sure there aren't contradictions (mostly conservation of mass), and then solves the 120-parameter differential equation. Then it does it 1000 times and collates the data (although I'm sure python could do that).

Mathematica seems like a much cooler language than Python. Python is like a dynamic Lisp but with an emphasis on imperative programming.

Mathematica is also a lot like Lisp and Mathematica also supports imperative programming. It's also a lot faster than Python.

In fact, I would argue that the only drawback with Mathematica is the fact that you become locked in. (whether or not you care about that is your prerogative.)

What do you mean by "dynamic Lisp"?

I guess I should have been clearer. It has a focus on run-time replacement of properties on objects. Duck-typing with an object-oriented emphasis. It is a dynamic language that supports lexical scoping but does not focus on the classic functional programming stuff.