August 19, 2010

How I wrote about P vs NP and lived to tell about it

I’ve known about the P versus NP problem” in computer science for a few years. It sits right at the intersection (which I love) of weird math, technology, and philosophy. So when a new proof” of the conjecture bubbled up into the news, I leapt at the opportunity to write a story about it for Technology Review.

The question I was curious to answer to was: What does P vs NP mean for the rest of us?

If it were proven one way or another—a very unlikely prospect—how would it affect the daily business of computing, if at all? I assumed that this was an easy” angle to take. Boy was I wrong. They don’t call it complexity theory” for nothing, after all. But while the technical details could choke a Vulcan, the distinctions between the various species” of computational problems (e.g., P,” NP,” NP-complete,” etc.) can seem quite intuitive at first. That’s what’s so intriguing about P-vs-NP: gobbledygook like this can actually map to very layman-friendly concepts, like playing Sudoku or arranging the seating assignments for a wedding reception.

And there’s the danger, of course. I’m used to trusting my own nose for an intuitive analogy. But when I finished my draft—and even after one of my sources vetted it for accuracy—my science-writer Spidey sense” was still subtly tingling. Thank god Scott Aaronson was willing to double-check it, because he uncovered half a dozen subtly-wrong-but-still-just-plain-wrong characterizations of basic concepts. I corrected them, and then asked him to triple-check. He found a couple more errors. Everything worked out and my editor was pleased in the end, but it was a very valuable reminder: when feeling extra-curious about something, be extra-careful.

The bright side is that this reporting experience only makes me more eager to write about P versus NP.” Maybe I’m a science-writing version of an adrenaline junkie: I feel like I barely made it out of this topic alive, but it’s just so damn interesting, and there’s so much more dangerous territory” to explore and bring back the goods from, that I can’t wait to go back.

[Postscript: I forgot to mention the_ huge importance of having a good editor on this story, or any one like it. Instead of fantasizing about murdering me, Will Knight _at Tech Review whipped my sorry excuse for a news lede into shape, worked late on several drafts, and actually thanked me in the end.]

Previous post
3 ways that curiosity helps me be a more productive freelancer This is a modified version of  a comment I wrote on Ed Yong’s “On the origins of science writers” blog thread, which is really fun and you should
Next post
What I learned in 2010 about making a living: everything is generative. I want to get this down, even if it’s only for myself. The most important lesson I learned about successful, productive freelancing in 2010 was