When you are performing research, there is a tradeoff. You start out trying to investigate a problem, but you end up mostly looking where the light is.
Research extends existing knowledge. You might be looking for your keys: but we have no flashlights, so you are forced to look where the light is. Perhaps you end up finding glass bottles instead of your keys! Well, then you write a paper on glass bottles, and help brighten up the area. You may hope to find your keys eventually, but your research has to look where the light is.
Right now, I hope to correlate long term trends in the Internet with policy decisions, so that policy makers will be able to rectify any changes that look bad some years after the fact, and so that future policy decisions will be made based on data rather than competing ideologies.
Unfortunately, we are still developing the tools to handle huge graphs, both in the sense of the software being unwritten, and in the sense that the theory remains somewhat of a mystery. So I’m looking in the light – I’m extending current tools to analyza graph stability, efficiency, and other properties, with the eventual goal of finally taking these tools I’m developing and utilizing them to effect change. That’s what I want to end up doing, but I have to look in the light if I want to find anything at all.
(from advice by Michal Young)
