PyDev of the Week: John Cook

This week we welcome John Cook as our PyDev of the Week! John has a fun Python blog that I read from to time and he graciously accepted my offer of interviewing him this week. Let’s take a few moments to get to know him better.

Can you tell us a little about yourself (hobbies, education, etc):

I’m a consultant working in the overlap of math, data analysis, and software development. Most projects I do have two of these elements if not all three. I had a variety of jobs before starting my own company, and most of them involved some combination of math and software development.


Why did you start using Python?

Several people suggested I try Python. I’d take a look at it and think “This doesn’t look so special. I don’t see what other people see in it.”  A few months would go by, then someone else would suggest I try Python, and I’d have the same reaction.
The thing that broke the cycle was a comment in an article by Bruce Eckel. He said something like “I teach Java but I can’t remember how to open a file in Java. But I can remember how to do it in Python.”  I realized then that when I’d looked at Python and thought it was kinda dull, that’s a good thing! Python doesn’t have flashy syntax. It’s kinda plain, in a good way.
I noticed later that while fans of other languages might want to show off some clever piece of code, you don’t see that much in the Python community. Instead you see people showing off what they were able to build with Python.
Even though I now mostly use Python for scientific computing, I didn’t know about Python’s scientific libraries when I started using it. I only used Python for the kinds of things I’d previously done in Perl, things like text munging. I was amazed when I found out later that so much mathematical code was available for Python.

What other programming languages do you know and which is your favorite?

I’ve written a lot of C++. When Python isn’t fast enough, I turn to C++, though I don’t do that often. In the last few years I’ve used R, C#, and Haskell on different projects.

I really like the consistency and predictability of Mathematica, though I haven’t used it in a while. I now use Python for the kinds of work I used to do in Mathematica. Even though some things are easier to do Mathematica, it’s worth some extra effort to keep from having to switch contexts and use two languages and environments. And of course Mathematica is expensive. Even if I decide the price of a Mathematica license is worth it for my own use, I can’t ask clients to buy Mathematica licenses.


What projects are you working on now?

I just started working on a big project involving signal processing and acoustics.


Which Python libraries are your favorite (core or 3rd party)?

I use SciPy daily. It’s my favorite in the sense that I depend on it and I’m grateful for the tremendous effort that has gone into it. I can’t say it’s my favorite in terms of API design; I wish it were more consistent and predictable.

I wish I knew pandas and SymPy better. I use them occasionally, but not often enough to keep their syntax in my head.

Conda is a sort of meta library rather than a library per se, but I really appreciate conda. It’s made it so much easier to install packages. I go back and forth between Windows and Linux, and it’s so nice to be able to count on the same libraries in both environments. Before some packages would install smoothly on one OS but not the other.


Where do you see Python going as a programming language? 

I don’t know much about where the core Python language is going, but it’s clear where the scientific computing stack on Python is going. The component libraries have come from independent projects, but a lot of work is going into integrating them into a coherent stack. I wrote a short blog post on this here.
Thanks so much!