Foreword

Forget the jokes about tasty snake dishes, here's the Python Cookbook! Python's famous comedian namesakes would have known exactly what to do with this title: recipes for crunchy frog, spring surprise, and, of course, blancmange (or was that a tennis-playing alien?). The not-quite-so-famous-yet Python programming community has filled in the details a little differently: we like to have fun here as much as the next person, but we're not into killing halibuts, especially not if their first name is Eric.

So what exactly is a Python cookbook? It's a collection of recipes for Python programmers, contributed by Python community members. The original contributions were made through a web site set up by ActiveState, from which a selection was made by editors Alex Martelli and David Ascher. Other Python luminaries such as Fredrik Lundh, Paul Dubois, and Tim Peters were asked to write chapter introductions.

Few cookbooks teach how to cook, and this one is no exception: we assume that you're familiar with programming in Python. But most of these recipes don't require that you be an expert programmer, either, nor an expert in Python (though we've sprinkled a few hard ones throughout just to give the gurus something to watch for). And while these recipes don't teach Python programming basics, most were selected because they teach something�for example, performance tips, advanced techniques, explanations of dark corners of the language, warnings about common pitfalls, and even suggestions that seem to go against accepted wisdom.

Most recipes are short enough for the attention span of the average Python programmer. For easy access, they are grouped into chapters, which contain either recipes for a specific application area, such as network programming or XML, or are about specific programming techniques, such as searching and sorting or object-oriented programming. While there's some logical progression among the chapters and among the recipes in a chapter, we expect that most readers will sample the recipes at random or based on the job at hand (just as you would choose a food recipe based upon your appetite or the contents of your refrigerator).

All in all, the breadth and depth of this collection are impressive. This is a testimony to Python's wide range of application areas, but also to its user community. When I created the first version of Python, more than 12 years ago now, all I wanted was a language that would let me write system-administration scripts in less time. (Oh, and I wanted it to be elegant, too.) I never could have guessed most of the application areas where Python is currently the language of choice for many�and that's not just because the World Wide Web hadn't been invented yet. In many areas, code written by generous Python users is as important as Python's standard library: think of numeric algorithms, databases, and user interfaces, in which the number of third-party choices dwarfs Python's standard-library offerings, despite the language's reputation that it comes with "batteries included."

Python is an evolving language. This cookbook offers some recipes that work only with the latest Python version, and a few that have been made obsolete by recent Python versions. Don't think this means that Python has built-in obsolescence! Usually, these obsolete recipes work fine, and the code that uses them will continue to work in future Python versions. It's just that when you're irked by a roundabout way of expressing a particular idea in code, there's often a better way available in a newer Python version, and we'd like you to know about it. On the other hand, it's sometimes useful to know how to write code that works for several Python versions at once, without explicitly checking version numbers all the time. Some recipes touch upon this topic, as well.

The increase in size of the community has caused some growing pains. Now that the early adopters are already using Python, growth must come from luring more conservative users to the language. This is easy enough, as Python is a very friendly language, but it does present new challenges. For example, as a special case of Murphy's law, anything that can go wrong during the installation process will go wrong for someone, somewhere, and they won't be pleased. The new Python users are often not savvy enough to diagnose and correct problems themselves, so our solution has been to make the installer even more bulletproof than it already was.

The same holds for almost all aspects of the language: from the documentation and the error messages to the runtime's behavior in long-running servers, Python gets more user-testing than I ever bargained for. Of course, we also get more offers to help, so all in all, things are working out very nicely. What this means is that we've had to change some of our habits. You could say that the Python developer community is losing some of its innocence: we're no longer improving Python just for our own sake. Many hundreds of thousands of individual Python users are affected, and an ever-growing number of companies are using or selling software based on Python. For their benefit, we now issue strictly backward-compatible bug-fix releases for Python versions up to 2 years old, which are distinct from the feature-introducing major releases every 6 to 12 months.

Let me end on a different aspect of the community: the Python Software Foundation. After the failed experiments of the Python Software Activity and the Python Consortium, I believe we have finally found the proper legal form for a nonprofit organization focused on Python. Keeping a fairly low profile, the PSF is quietly becoming a safe haven for Python software, where no single individual or organization can hold a monopoly on Python, and where everybody benefits. The PSF, in turn, benefits from the sales of this book: a portion of the royalties goes to the PSF, representing the many Python programmers who contributed one or more recipes to the cookbook project. Long live the Python community!

�Guido van Rossum

Reston, Virginia

April 2002