Review of O'Reilly's Python Cookbook, Second Edition

Full disclosure: I purchased the first edition of this book, and received the second edition as a reviewers copy. I work for Zope Corporation.

The O’Reilly Python Cookbook is a fun resource for Python programmers at most levels. The fun comes in part from the personalities that shine through the introductions and the community-authored recipes gathered from the ActiveState Python Cookbook website. The other fun comes from the smorgasbord of topics and technologies laid out for the reader, encouraging browsing and experimentation. New Python programmers will find recipes that highlight some of the newer features of the Python language, and experienced Python programmers will likely find thought-provoking recipes both peripherally and directly related to their specialties and interests.

Like the first edition, this second edition covers a wide range of topics. Each topic has a usually-interesting introduction by well-known names in the Python community. Some topics are of general interest—shortcuts and algorithms, for instance—while others explore somewhat more specialized topics, such as networks, XML, and databases. Each cookbook recipe I read was impressively short, while often still having enough weight to them to address non-toy usages of the approaches. Many examples can also be used as introductions to the modules and packages they use. Another important similarity to the first edition is that a portion of the proceeds from the book sales are donated to the Python Software Foundation.

This edition of the cookbook does have some significant changes from the previous one. While the first edition addressed Python versions in the 1.x and 2.x line, this one addresses only Python 2.3 and 2.4. It adds about a third more recipes than the first edition, and tops the first edition’s page count by over 200 pages. Some of the new additions address relatively new packages, such as the datetime, dateutils, pytz, and decimal packages, while others focus on older packages that are waxing in their influence and popularity, such as Twisted. Many recipes are more polished, such as the “Synchronizing All Methods in an Object,” which has an arguably more general purpose approach and an interesting discussion in the new edition.

Omissions are sometimes as interesting as inclusions in recipe books, but an interpretation of their significance can be difficult. The new edition dropped almost 100 of the first edition’s recipes. In the “Distributed Programming” topic, the SOAP references in the first edition have disappeared, and a recipe for Twisted’s Perspective Broker has surfaced. While the Perspective Broker is an interesting technology with an elegant cookbook example, the seeming judgement—SOAP is now less worth discussion than one of the Twisted project’s packages—might raise some eyebrows. Similarly, Zope and the ZODB (Zope Object Database) are acknowledged as heavyweights in the introductions to their respective topics (“Web Programming” and “Persistence and Databases”). Zope 3 is even granted a rave review: “The new, revolutionary major release, Zope 3, makes Zope more Pythonic and powerful than ever.” But not a single recipe can be found for either, in either edition.

This returns us to the initial observation of this review: the cookbook is a community driven project, and thus reflects the personalities of the programmers who contributed to it, rather than necessarily to trends of the Python or IT community. But it also benefits from the energy and enthusiasm of the contributors who often have palpable excitement for the technologies they are demonstrating. This also makes me more interested in the book as a source for introductions than as a source for ready-made recipes.

If a picture is worth a thousand words, a working, explained example of a new package ought to be worth quite a few too. The cookbook is a fun browse and a fun introduction to other aspects of the Python language. While I have found myself using one or two of the recipes from the first edition, that use alone is not enough to justify the cover price. It is more compelling to me as a fun introduction to Python topics and approaches. In that light, I recommend it.