Beginning Python®: Using Python 2.6 and Python 3.1

Chapter 17. Extension Programming with C

Don't let anybody mislead you: well-written code in C will always execute faster than code written in Python. Having said that, don't be misled: Developing code in Python will always be faster than developing code in C.

This may seem like a dilemma at first. You want to have fast code, and you want to produce it quickly. Balancing these, and the problem it creates, is actually easily solved. Develop your code in Python. After all, developer's time is much more expensive than the computer's time. Plus, humans have a miserable track record of predicting where a bottleneck is going to occur in a system. Spending time optimizing code up front by doing things like taking a lot of time to write a new program in C is usually wasted time. This is what led the esteemed computer scientist, C. A. R. Hoare, to say, "Premature optimization is the root of all evil." Of course, he was only talking about computer programs, but the point is there.

If you've written your code, optimized your algorithms, and still find performance is unacceptable, you should profile your application by finding out where it's spending its time, determine where the bottlenecks are, and reimplement those small parts in C as a Python extension module. That's part of what this chapter is about.

Or if you already have an existing body of code written in C and you want to leverage that from within Python, you can create a small Python extension module exposing that C code to your Python ...