I should have known the lion by his claw… Martin Wattenberg is at it again. From Matt Jones’s excellent blog blackbeltjones, I found this fascinating report on recent work at IBM on something called history flow. History flow is a way to visualize the history of a living document. And the really nice touch here is that they chose to visualize the history of wiki articles in Wikipedia, particularly those that have sensitive topics like Islam, abortion, evolution, and Iraq. Not surprisingly, this last one grew dramatically as American and British troops prepared to attack.
Only after admiring the tremendous coolness of history flow for a while did I stumble across the fact that Martin was behind it. Matt Jones realized this a few days later, and added these thoughts, which will give you a little more information on Martin. About the same time, Clay Shirky, the ever thoughtful groupware pundit, rang in with a few thoughts of his own.
In a much more limited way, this is similar to work we did analyzing the entries of the MATLAB programming contest by watching who changed the code where, how much, and what difference it made. It would be fun to apply the history flow code to the contest database, particularly since we have a performance metric that the wiki prose lacks.
I was also reminded of Ben Fry’s haplotype plot, because after all, a genome is really just a kind of wiki that’s been subjected to natural selection. It seems only fitting that the history flow concept should apply here too. I know about some other fun haplotype visualization work done at the Whitehead by Gabriel, et al (including my friend John Higgins) last year: The Structure of Haplotype Blocks in the Human Genome. Science 296:2225-2229. I’d give you the link to it, but the Science site is down right now, and they hide the article behind a subscription barrier anyway. But it is certainly intriguing to consider the history flow of your genome extending back to your trilobite cousins. It’s a long story, but a hell of a read.