A Journey to Python
by Alex Martelli, 2006 recipient of the Frank Willison Award
After a quarter century of experience in programming (once as a sideline of my main job as a hardware designer, at Texas Instruments and IBM Research, but as my main job for over half that time, at IBM Research and Cad.Lab/think3), I finally stumbled upon Python in 1999. This was thanks to the kind nagging of a friend and colleague whose judgment I respected and to whom I still feel grateful for his insistence (Alessandro Bottoni).
I had plenty of experience with languages high and low, from microcoding (even FPGA "programming", lower-level yet :-), through assembly code for many machines, through standard "high-level" languages (Pascal, Fortran, PL/I, C, C++, Java...), all the way to "higher-level" languages which I had always used whenever feasible (Rexx, APL, AWK, Icon, Perl, Scheme...). But never had I met something that "fit my brain" as perfectly as Python.
At my "real job", at that time, the language du jour was mostly C++, with a smattering of Visual Basic and a proprietary scripting language that had been developed in-house and was embedded in the applications we developed and sold. However, in parallel, I was pursuing, at home, a research project on the mechanics of the game of bridge (an ancient but always-burning passion of mine) which resulted in articles published in January and February 2000 in the extremely prestigious magazine "The Bridge World" under the title "How Shape Influences Strength". Unfortunately, the large software base underpinning that research effort had grown "organically" into an unholy mix of C++, Perl, and home-grown languages (an open-source one for dealing and selecting bridge hands, one I had developed myself years before for bridge bidding under the uninspiring name of BBL for "Bridge Bidding Language", and more besides) that had grown essentially impossible to maintain and develop yet further, in order to continue my research. So I was looking for the best way to re-code the whole system from scratch. Python offered me THE way to do that.
I fell in love, and immediately started making a nuisance of myself on comp.lang.python, posting copiously (and, apparently, in ways that many people found interesting, fun, or helpful). Ominously, even my very first post contained the sentence "Books matter—a lot" (in the context of criticizing one existing book and praising two others).
A few months later (thanks to Greg Wilson reaching out to me), I was negotiating with O'Reilly about writing a Python book. I really wanted to do a "Python Cookbook", but O'Reilly had other ideas for that, in cooperation with ActiveState, based on an online cookbook; so I undertook instead to write Python in a Nutshell, and meanwhile I tried to help that cookbook project along by posting recipes, commenting on others' posts, and so on.
Smack in the middle of the Nutshell's gestation, an emergency emerged (as emergencies will :-)—the Python Cookbook just wasn't going to happen unless somebody with lots of Python nous AND some writing and editing skills could devote a very substantial slice of time to selecting and editing recipes. I suspended the Nutshell work, took all my accumulated vacation from work, and then some. (I worked in Italy, accumulating vacation at Italian rates, but typically took amounts of vacation on more of an American scale. So, after a few years on the job, I had QUITE a lot of accumulated vacation. And in Italy, by law, you never "lose" vacation days you're due.) I threw myself into the task; the first edition of the Cookbook emerged from that long and frantic effort. ActiveState's people who worked with me on that project, chiefly David Ascher who's credited as co-author of the Cookbook, expressed their appreciation of my work by granting me the "Activators' Award" in 2002 at OSCON.
By the time that was done, I was even deeper in love with Python; while also going back to writing the Nutshell in my spare time, I redoubled my efforts at work to "muscle in" the use of Python alongside the "corporately blessed" languages. I did find a point of leverage, using Python's win32 extensions to implement COM objects for Windows interfacing to legacy functionality in our apps—very much of a "skunkworks" effort without any real support from top management, who'd have preferred "Microsoft-blessed" languages. Ironic, if you think that today, finally!, Microsoft's IronPython is a first-rate implementation of Python on .NET... Anyway, Python is now firmly ensconced at that firm, since so many apps have been developed and sold based on those COM objects and their later dressing up in .NET :-). But at the time the experience was very frustrating and brought home to me how much the company had changed since I had joined it many years before when it was barely out of the startup stage—grown so fast (perhaps too fast), acquired "professional" top management that didn't think any more that it needed to listen to input from "top techies"/"highly technical managers" such as myself.
Right about then, I got an offer to consult as a freelance Python specialist for a Swedish startup, and I took the opportunity, quitting my "safe job" and switching to freelance work, mostly but not exclusively for that Swedish company. It DID mean a lot of "commuting by plane", sigh, but it also gave me the opportunity to work almost exclusively in Python (and in C, coding extensions for Python). I also managed to offer some contributions to Python itself, and develop the gmpy project to offer advanced multiprecision arithmetic, number-theoretical functions, etc., on top of the LGPL-licensed GMP library (some of those advanced functionalities are precious to my ongoing work in bridge simulation and analysis, though by now my ongoing work maintaining and developing gmpy is mostly something I see as a contribution to the community).
I did complete the Nutshell, wrote many Python articles for various magazines, both Italian and US ones, and went on a binge doing presentations at all sorts of conference, such as PythonUK, EuroPython, PyCon, OSCON, and many Italian open-source ones.
It's also through Python that I got back in touch with a woman, Anna, whom I had known years before but had lost contact with. She was studying Python and recognizing my name mailed me to ask me for some help. She's now my wife, and became the first woman member of the PSF. We did the second edition of the Python Cookbook together, and although she's not officially an author of the second edition of Python in a Nutshell, she was so incredibly helpful with it that she might as well be.
Through my Python books I got in touch with Google, at the time a small company but already a substantial customer for the Nutshell. I tried to sell them on my services as a consultant and found myself with an excellent job offer instead. So I dropped the freelancer's life, moved to California, and here I happily live, serving as "Uber Tech Lead" for that deep-infrastructure software we call "Production Systems", where Python plays a crucial role (though not an exclusive one, as C++ is also important here).
I was pleasantly surprised to find that working at Google, even in such a fascinatigly challenging position as Uber Tech Lead is, allows enough "work-life balance" (or in my case work-work balance :-) to let me write the second edition of the Nutshell (thanks also to Anna's already-mentioned absolutely outstanding support). Ever since I finished that task, I've devoted my time to more and more promotion of Python, and best Python practices, style, and usage, both at Google and everywhere else. My dream is to write another book, on Python idioms, design patterns, and development methods, but I think I'll take one more year off bookwriting before I start doing that. In the meantime, I'm also writing, presenting and teaching about "highly technical management of software development", a subject on which many myths exist and on which I have substantial personal experience and reflection to share with others.