Complex numbers are important in so many aspects of math and physics, and despite the name they are not so complex. This book has a lot to teach even those who think they know complex numbers well, since most of us never learn much about the geometry of these numbers. And for those new to the subject, this is an endlessly stimulating introduction. available here at amazon: http://www.amazon.com/Visual-Complex-Analysis-Tristan-Needham/dp/0198534469 also see the author's page about the book here: http://www.usfca.edu/vca/ Also, I'll throw in the Feynman Lectures on Computation, since it is a nice introduction to the physics of computing; plus it's hard to go wrong with anything by Feynman.

In my lab at UCSB (http://www.physics.ucsb.edu/~martinisgroup) we have developed a package called LabRAD (http://sourceforge.net/projects/labrad/) for doing our data acquisition and experiment control. It is a network protocol that gives us network transparency as well as language independence. We have modules for talking to serial ports, GPIB, raw ethernet, etc, as well as servers that use these busses to talk to the various instruments in the lab. The backbone is written in Delphi, but most of the modules are written in python, which is also our scripting language of choice for running everything, and for data analysis. I can only agree with previous posts about the general awesomeness of python/numpy/scipy and matplotlib for plotting. The network transparency is great because we can monitor the whole lab and run experiments from anywhere. This is Santa Barbara after all, and nothing beats doing physics while lounging on the beach with a laptop!