Credit: David Ascher, ActiveState, co-author of Learning Python (O’Reilly)
Programming languages are like natural languages. Each has a set of qualities that polyglots generally agree on as characteristics of the language. Russian and French are often admired for their lyricism, while English is more often cited for its precision and dynamism: unlike the Académie-defined French language, the English language routinely grows words to suit its speakers’ needs, such as “carjacking,” “earwitness,” “snail mail,” “email,” “googlewhacking,” and “blogging.” In the world of computer languages, Perl is well known for its many degrees of freedom: TMTOWTDI (There’s More Than One Way To Do It) is one of the mantras of the Perl programmer. Conciseness is also seen as a strong virtue in the Perl and APL communities. In contrast, as you’ll see in many of the discussions of recipes throughout this volume, Python programmers often express their belief in the value of clarity and elegance. As a well-known Perl hacker once said, Python’s prettier, but Perl is more fun. I agree with him that Python does have a strong (as in well-defined) aesthetic, while Perl has more of a sense of humor. I still have more fun coding in Python, though.
The reason I bring up these seemingly irrelevant bits at the beginning of this book is that the recipes you see in this first chapter are directly related to Python’s aesthetic and social dynamics. In most of the recipes in this chapter, the author presents a single elegant language feature, but one that he feels is underappreciated. Much like I, a proud resident of Vancouver, will go out of my way to show tourists the really neat things about the city, from the parks to the beaches to the mountains, a Python user will seek out friends and colleagues and say, “You gotta see this!” Programming in Python, in my mind, is a shared social pleasure, not all that competitive. There’s great pleasure in learning a new feature and appreciating its design, elegance, and judicious use, and there’s a twin pleasure in teaching another or another thousand about that feature.
When we identified the recipe categories for this collection, our driving notion was that there would be recipes of various kinds, each aiming to achieve something specific—a souffle recipe, a tart recipe, an osso buco recipe. Those would naturally bunch into fairly typical categories, such as desserts, appetizers, and meat dishes, or their perhaps less appetizing, nonmetaphorical equivalents, such as files, algorithms, and so on. So we picked a list of categories, added the categories to the Zope site used to collect recipes, and opened the floodgates.
Pretty soon, it became clear that some submissions were really hard to fit into the predetermined categories. These recipes are the Pythonic equivalent of making a roux (melted butter or fat combined with flour, used in sauce-making, for those of you without an Italian sauce background), kneading dough, flouring, flipping a pan’s contents, blanching, and the myriad other tricks that any accomplished cook knows, but that you won’t find in any “straight” recipe book. Many of these tricks and techniques are used in preparing various kinds of meals, but it’s hard to pigeonhole them as relevant for a given type of dish. And if you’re a novice cook looking up a fancy recipe, you’re likely to get frustrated quickly, as these techniques are typically found only in books like Cooking for Divorced Middle-Aged Men. We didn’t want to exclude this precious category from this book, so a new category was born. That explains why this chapter exists.
This chapter is pretty flimsy, though, in that while the title refers to shortcuts, there is nothing here like what one could have expected had the language in question been Python’s venerable cousin, Perl. If this had been a community-authored Perl cookbook, entries in this category would probably have outnumbered those in most other chapters. That is because Perl’s syntax provides, proudly, many ways to do pretty much anything. Furthermore, each way is “tricky” in a good way: the writer gets a little thrill out of exploiting an odd corner of the language. That chapter would be impressive, and competitive, and fun. Python programmers just don’t get to have that kind of fun on that kind of scale (by which I mean the scale of syntactic shortcuts and semantic-edge cases). No one gives multi-hour talks about tricks of the Python grand masters... Python grand masters simply don’t have that many frequently used tricks up their sleeves!
I believe that the recipes in this chapter are among the most
time-sensitive of the recipes in this volume. That’s
because the aspects of the language that people consider shortcuts or
noteworthy techniques seem to be relatively straightforward,
idiomatic applications of recent language features. List
zip, and dictionary methods such
setdefault are all relatively recent additions
to the language, dating from Python 2.0 or later. In fact, many of
these newish language features were added to Python to eliminate the
need for what used to be fancy recipes.
My favorite recent language features are list comprehensions and the new
applicability of the
tokens to function calls as well as to function definitions. List
comprehensions have clearly become wildly successful, if the authors
of this volume are representative of the Python community at large,
and have largely demoted the
filter built-in functions. Less powerful, but
equally elegant, are
**. Since Python
2.0, the oft-quoted recipe:
def method(self, argument, *args, **kw): # Do something with argument callable(*args, **kw)
can now be done much more elegantly as:
def method(self, argument, *args, **kw): # Do something with argument callable_object(*args, **kw)
apply built-in function is still somewhat
useful, at least occasionally, but these new syntactic forms are
elegant and provably Pythonic. This leads me to my closing comment on
language shortcuts: the best source of shortcuts and language tricks
is probably the list of language changes that comes with each Python
release. Special thanks should be extended to Andrew Kuchling for
publishing a list of “What’s
new with Python 2.x,” available at http://amk.ca/python/, for each major release
since 2.0. It’s the place I head for when I want a
clear and concise view of Python’s recent evolution.