In case you missed it, there is a lovely interview with Donald Knuth over at the Artificial Intelligence Podcast. I subscribe to it in my podcast reader, so I listened to it…but I’ve embedded the video below if you prefer to watch interviews. There’s something quite intimate about listening to someone’s voice while riding a bus, or walking down the street…and Knuth’s humble, meandering way of talking and telling stories was a real pleasure to listen to.
I actually haven’t finished the whole thing yet, but I’ve been struck by how important writing is to his work. If you are familiar with literate programming this should really come as no surprise. But hearing him talk about writing (and reading) on equal terms with researching programming was really quite remarkable.
At one point (33:56) Knuth recalls how he began writing The Art of Computer Programming in 1962, and effectively became a what he calls a journalist of algorithms:
In 1962 when I started writing down a table of contents, [The Art of Computer Programming] wasn’t going to be a book about computer programming in general. It was going to be a book about how to write compilers. I was asked to write a book explaining how to write a compiler.
At that time there were only a few dozen people in the world who had written compilers and I happened to be one of them. I also had some experience writing, for the campus newspaper and things like that.
So I said okay great, I’m the only person I know who has written a compiler, but hasn’t invented any new techniques for writing compilers, and all the other people I knew had super ideas…but I couldn’t see that they would be able to write a book that would describe anybody else’s ideas with their own. So I could be the journalist, and I could explain what all these cool ideas about compiler writing were.
I have by no means read the entirety of TAOCP, I use it mostly as a reference book. But one thing that is striking to anyone that dips into its pages is the historical depth of his explanations and citations. I honestly think of Knuth as more a historian of algorithms than a journalist…but the contemporary nature of what he was writing about does lend itself to that mode of writing I suppose.