The Complex Systems Reading Club is having a go a paper on power laws by Mark Newman this week. This should be good. A good false grasp of power laws looks to be one of the things between me and that sociologist’s understanding of complexities studies that I’m trying to achieve. These laws are spoken of in common company as though we all understand them, and clearly we do, except me. Near as I can tell, the idea seems to be that nature beat Chris Anderson to the punch. A discussion of a review paper on the topic over ketchupy french fries will probably help me out here. And this Mark Newman seems to be a clever cat.

(Aside: the School of Information had a host of wicked cool speakers in last month including Chris Anderson. But for reasons not at all clear to me, they had them all speak during the fall break, when everyone with sense was somewhere else. Perhaps they don’t have the information?)

After the power law paper was distributed Mark Newman pointed out that he had “made a lot of updates and corrections to this paper since it was first published” and that the better version was here.

This is I think the first time I’ve used a paper from arXiv (maybe as I venture deeper into the statistical physics this will become more common). I’m still all crushy on the idea of the gradual transmogrification of peer-reviewed scientific publishing into something more organic and transparent. What hadn’t occurred to me previously about the arXiv was that it not only allows for a paper to be born piece-meal and iteratively, but by existing in an editable space the paper never really has to be finished. Potentially, the most canonical thing about it is just the URL it’s attached to, and every time you needed the paper you would access it fresh and perhaps perform a “compare” to see what new has developed since the last time you read it. There will be a continuum of rigor vs. freshness with wikis and inked napkins on one end and text books on the other and academic papers will perhaps shift within it, or have some way of indicating where in that continuum they claim to exist. Papers will have tiers of authors, or just an ‘author cloud’. Every paper about networks will ironically be more a network than a paper. Bruce Sterling will dress in rags and stand on campus orating in Austin-tinged Texan accents about the coming infoapocalype, again.

Of course this raises a slew of important questions about how the goals of peer-review are alternatively to be accomplished if there is no One True Version of the paper to be signed off on. You trust a paper from a respected journal because the reviewers are willing to underwrite the contents with their own credibility. But they won’t tie their names and reputations to a paper that may change to say something less valid 6 months from now. So how do we do it?

I sure don’t know but I’m real convinced that it can be done, and probably at least as well as in the conventional system. I have this faith because among other things, it seems to be being done. Maybe the people at the School of Information can tell me how.