Guido van Rossum (1956) is the founding father of the Python programming language, one of the most popular development tools in the world. On 21 November 2019 CWI will award him the Dijkstra Fellowship.
What led you to come up with a brand new programming language during your time at CWI?
“I started at CWI as a junior programmer on a research team with Lambert Meertens, Leo Geurts and Steven Pemberton. They wanted to develop a language which would enable people without programming experience – such as scientists – to start writing computer programs fairly quickly.”
“It was at the time that Basic was on the rise due to the arrival of the microcomputer. Meertens looked at this inadequate language with horror. ‘Stamp out Basic!’ Was his motto. In the end, ABC, as our language was called, would not work. The target group could not use it on their microcomputers, which were not powerful enough for it, while Unix users already had other tools. Those users thought ABC was an odd man out.”
“Then I came across the so-called Amoeba project. That was a distributed operating system based on a microkernel, developed by Andrew Tanenbaum at the Vrije Universiteit Amsterdam and Sape Mullender at CWI. Not aiming at popularizing their operating system, their first and foremost goal was writing papers. Scientifically it was a breakthrough indeed: those papers are still being studied. I myself was not a researcher but a programmer on that project. I must say though that there was an atmosphere at CWI in which programmers had a major input in the projects.”
When did you start developing Python?
“During the Amoeba project I realized that if we had to write all the required applications in the C programming language, it would be very time consuming. Therefore I started to wonder if I could write an improved version of ABC.”
To what extent was CWI a stimulating environment for this?
“The colleagues, of course, but also the enormous freedom to learn from anybody about everything. That learning was often done by taking up projects without knowing exactly what the outcome would be. That freedom was simply great.”
And then Python caught on in the US!
“The timing was much better now. When we developed ABC, the internet didn't exist yet. If you wanted to send a piece of code to the US, you had to board a plane to deliver a magnetic tape in person. But with Python I had Usenet at my disposal – not really the internet yet, but you could send source code over it. That, of course, has contributed enormously to its spread.”
You were then offered a full position as a researcher at CWI. You turned it down though.
“That’s right. I was already in my mid-thirties. Meanwhile, Python began to strike in the US. I thought it would be much more fun to get actively involved. Fortunately I realized on time that I’d rather write code than scientific articles.”
You have been living in the US for almost 25 years now. How did your oversees career lift-off?
“At one point, one of my Python contacts on Usenet asked me if I wanted to come to the US for a month to give lectures and participate in discussions about Python. I met two people with jobs, who tried to convince their bosses that it would be beneficial to hire me. That's how I ended up at the Corporation for National Research Initiatives in Virginia, a company that at the time aimed at basing all their development work on Python.”
“After a few years the start-up fever caught me. I worked for a number of small companies, none of which turned out to be a success. I came to Google in 2005. They were no longer a start-up at that moment: at least ten thousand people were working there.”
Didn't your Dutch friends find that really cool?
“They were very impressed indeed – Google was very hot. For me it was also special. For the first time in my life I worked at a large and successful company.”
A famous company that is, which also happened to use your language.
“Yes, otherwise I wouldn't have been fun. I created a tool to improve the code review process. The practice was that when a programmer made something, another programmer had to check it manually. I saw room for improvement there. Soon I had gathered a whole team around me. The resulting tool was characterized by white text with green and red blocks in it. When I walked through the corridors, or visited another branch, I saw those characteristic screens everywhere. That was a special experience, I must say.”
Yet you didn’t stay.
“No. I went to Dropbox. In terms of corporate culture, that was kind of a smaller Google where you got to know more people. I made all kinds of Python tools for six and a half years. I also led the conversion from Python 2 to Python 3.”
You recently retired. But it seems safe to bet that your brainchild will keep you occupied.
“I stepped down as ‘Benevolent Dictator’, as I was called. But I still think I will have to answer a lot of email. I hope to be able to limit that to four hours a day.”
Advice for a new generation
What would you advise to a new generation? Which parts of the software realm will be the hottest in the years to come?
“I especially hope that they will make as many different choices as possible. After all, not everyone can earn a living as a developer of programming languages, or as an AI expert. Anyone who makes sufficient efforts can find interesting work in all sorts of areas – as long as they are sufficiently talented, of course.”
What is the biggest challenge they will face?
“The most important problem of our time is that computer systems are so incredibly distributed. All kinds of applications use a huge complex of computers that differ in size, processing power and operating system. That causes a lot of headaches. I actually think it is a miracle that this complex system works at all.”
How do you feel about receiving the Dijkstra Fellowship?
“I’m as proud as you can imagine!”
By: Ed Croonenberg