Re:Paul Krugman: How I work
by amix on Tuesday October 14, 2008 @07:17AM
(#332131)
I think his way of doing research is the way computer scientists do research. I.e. instead of taking a very complex programming model (like Java's or Haskell's), most CS researchers invent their own mini languages (that in a lot of cases are simple extensions of lambda calculus or turning machines), these mini languages makes it much easier to reason about specific things. An example: Java's generics were first presented in a paper that added them to a mini-language that only mimicked some properties of Java.
That said, his way of working also seems to be very "hackerish", i.e. instead of using "the normal ways", he hacks a model together that show some properties that the conventional and more complex models have a hard time showing. This is basically also how great hackers work, e.g. Paul Graham's Arc (it's very unconventional and simplified - but great at showing the essence of the hacker's ideas).
That said, his way of working also seems to be very "hackerish", i.e. instead of using "the normal ways", he hacks a model together that show some properties that the conventional and more complex models have a hard time showing. This is basically also how great hackers work, e.g. Paul Graham's Arc (it's very unconventional and simplified - but great at showing the essence of the hacker's ideas).
Anyway, great link!