It's common in math and computer science for people to prove important theorems sort of in passing, on the way to some other result. At least, it looks that way to an outsider-- Fermat's Theorem and the Poincare Conjecture are the high-profile examples that come to mind.
In that spirit, Scott Aaronson helpfully distills all of academic research into two paragraphs in the course of making a handy Frequently Asked Questions list for people who might want to hire him:
I know I'm not supposed to say this in an interview, but I don't have a vision. I have this annoying open problem, that conjecture, this claim that seems wrong to me. I know some people have a coherent vision for where their research is headed. And in experimental areas, obviously you have to justify what you're going to do with your $200 million of equipment. But at least in theoretical computer science, having a "vision" always seemed incredibly difficult to me.
For example, let's say you have a vision that you're going to solve problem X using techniques A, B, C. Then what do you do when you learn that techniques A and C are nonstarters -- but that technique B, while it's useless for X, does solve a completely unrelated problem Y? What you do is make up a story about how Y was the problem you wanted to solve all along! We all do that: drawing targets around where the arrows hit is simply the business we're in.
"Drawing targets around where the arrows hit" would be a great tagline for a blog.
- Log in to post comments
Astronomy is ripe for "arrow" jokes. e.g., "Of course the star with the planet was easy to find! There's an arrow pointing right at it!"
mollishka: Oddly enough, last Thursday I attended a lecture where Geoff Marcy (dude who finds planets) made exactly that joke in that exact context (planet finding)