SciPy 2009

In approximately 10 hours I will be in the air heading to Caltech for the 2009 SciPy Conference. Thanks to Randy and Kyle for allowing me to tag along and learn first-hand about the current state of Python-based mathematical and scientific programming technologies.

The week is broken down into three sections. On Tuesday and Wednesday I'll be attending "Tutorials". These are , in a sense, a series of workshops where the creators of these technologies teach the attendees about how to use their products as well as share some advanced tips and tricks and information on contributing to the development of their projects. (These are all open-source programs.) In particular, I'll be attending the "advanced" tutorials track where I'll be learning about the following:

• Advanced topics in numpy: numpy offers a lot of numerics capabilities to the Python library. Some describe it as making the Python programming language more "Matlab-like".
• Advanced topics in matplotlib: matplotlib is a package for producing nice-looking plots. Sage and Clawpack use it for its 2D plotting routines.
• Symbolic computing with sympy: sympy is, in a way, the opposite of numpy. It specializes in symbolic manipulation rather than numerical computation. Again, Sage uses it for a good portion of it's symbolics and expression handling.
• Statistics with scipy: scipy, the creators of which are hosting the conference, provides tools for scientific application. It ties in very well with numpy.
• Cython: Cython is a tool for easily creating "C extensions". These are, essentially, C programs that are designed to have a Python interface. It allows the possibility of writing some very fast Python code.
• Using GPUs with PyCUDA: programming GPUs is a hot topic right now in scientific computing. PyCUDA allows you to write Python programs that will take advantage of your GPU's superior processing power.
• Designing scientific interfaces with Traits: Traits is a package that provides advanced Python object attribute handling among other things. Part of Traits allows you to create GUIs for your programs.
• Mayavi: Mayavi provides easy and interactive visualization of 3D data.

Again, the great thing about these technologies is that they're all open-source. I now have them all installed on my computer and am ready to learn how to use them!

That's the tutorial section of the conference. Thursday and Friday host the core of the conference which are the talks. The Keynote will be given by Peter Norvig, Director of Research at Google. It's titled: "What to demand from a Scientific Computing Language -- Even if you don't care about computing or languages". Kyle will be giving a talk about the use of Python in Pyclaw. If you're interested to see what the talks will be about, visit the SciPy Conference Schedule.

The Conference wraps up with "coding sprints" on Saturday and Sunday. To put in bluntly, a coding sprint is when a whole bunch of developers get together in the same, big room and code all day. What's nice about it is that if you have a question about using a particular technology chances are that an expert is sitting somewhere in the room.

At the very least, this week will do well to output a lot of Clawpack-related work. I hope to have a refined Sage implementation of Clawpack up and running by then end of it all. I also look forward to seeing the Cython developers' progress on this summer's Fortran + Cython / f2py modifications, since it may prove to be useful in the development of Clawpack 5.0.