Nov 25 2008
With enough examples, all problems are familiar
I have been thinking about a particular problem now for about two years. The problem doesn’t really matter — it’s a computer program that I want to write one day, when I get some time and space to do so.
What does matter is that last week I realised, almost completely out of the blue, that my long-standing problem was very different in nature to what I had assumed. Not only that, but as soon as I realised this the problem became massively more tractable. It was like turning an unknown corner and finding yourself in a familiar neighbourhood: it all just fell into place.
It’s very comforting to know that the fundamental problem I am trying to solve has been solved many times before, and there is a large body of research into the different approaches to take. The surface problem is still interesting, and to the best of my knowledge no-one has done anything similar before. But it’s new in the sense that the latest model of car is new: the problem domain is well mapped out, and all that matters is the details.
“EU bus regulations are NP hard … so they use Haskell to enforce bus timetable correctness ” http://www.haskell.org/communities/11-2008/html/report.html#sect7.7
Did you notice I was also on that page too?
Ha, no I didn’t! You’re awesome!