Preface
Python is rightfully viewed as a general purpose language, well suited for web development, system administration, and general purpose business applications. It's has earned this reputation well by powering web sites such as YouTube, installation tools integral to Red Hat's operating system, and large corporate IT systems from cloud cluster management to investment banking. Python has also established itself firmly in the world of scientific computing covering a wide range of applications from seismic processing for oil exploration to quantum physics. This breadth of applicability is significant because these seemingly disparate uses often overlap in important ways. Applications that can easily connect to databases publish information to the web, and efficiently carry out complex calculations are now critical in many industries. Python's primary strength is that it allows developers to build such tools quickly.
Python's scientific computing roots actually go quite deep. Guido van Rossum created the language while at CWI, the Center for Mathematics and Computer Science, in the Netherlands. As interest developed outside the center, others began to contribute. The first several Python workshops, starting in 1994, were held an ocean away at scientific institutions such as NIST (National Institute of Instruments and Technology), the US Geological Society, and LLNL (Lawrence Livermore National Laboratories), all science centric institutions. At the time, Python 1.0 had recently been released and the attendees were just beginning to hammer out the design of its mathematical tools. A decade and a half later, it is gratifying to see how far we have come both in the amazing capabilities of the tool set and the diversity of the community. It is somehow fitting that the first comprehensive book (that I know of) covering the primary scientific computing tools for Python is composed and published, another ocean away, in Chinese. Looking forward a decade and a half, I can hardly wait to see what we will all build together.
Guido, himself, was not a scientist or engineer. He sat squarely in the computer science branch of CWI and created Python to ease the pain of building system administration tools for the Amoeba operating system. At the time, the tools were being written in C. Python was to be the tool that“bridged the gap between shell scripting and C.”Operating system tools are not even in the same neighborhood as matrix inversions or fast Fourier transforms, but, as the language emerged, scientists around the world were some of its earliest adopters. Guido had succeeded in creating an elegantly expressive language that coupled nicely with their existing C and Fortran code. And, in Guido, they had a language designer willing to listen and add critical features, such as complex numbers, specifically for the scientific community. With the creation of Numeric, the precursor to NumPy, Python gained a fast and powerful number crunching tool that solidified Python's role as a leading computational language in the coming decades.
For some, the term“scientific programming”conjures up visions of intricate algorithms described from“Numerical Recipes in C”or forged in late night programming sessions by graduate students. But the reality is the domain encompasses a much wider range of programming tasks from low level algorithms to GUI development with advanced graphics. This latter topic is too often underestimated in terms of importance and effort. Fortunately, Ruoyu Zhang has done us the service of covering all facets of the scientific programming in this book. Beginning with the foundational Numpy library the algorithmic toolboxes in SciPy he provides the fundamental tools for any scientific application. He then aptly covers the 2D plotting and 3D visualization libraries provided by matplotlib, chaco, and mayavi. Application and GUI development with Traits and Traits UI, and coupling to legacy C libraries through Cython, Weave, ctypes, and SWIG are well covered as well. These core tools are rounded out by coverage of symbolic mathematics with SymPy and various other useful topics.
It's truly gratifying to see all of these topics aggregated into a single volume. It provides a one-stop shop that can lead you from the beginning steps to a polished and full featured application for analysis and simulation.
Eric Jones
2011/12/8