: New: Book Report: The Reputation Society

I'm not game control for every game out there. E.g., I'm not on Game Control for this game; I was just passing along word of its existence after I saw it mentioned on a flyer. But I can see why you might think I was in game control. I have a reputation, and a reputation's a strange thing. Especially when you try to smoosh it into a computer… as discussed in The Reputation Society.

It's a book of papers about reputation in all of these newfangled social networks we're building: Slashdot karma, Amazon's Top Customer Reviewers, folks on a StackOverflow leaderboard… any time some algorithm tries to figure out a recommendation of people a la maybe you don't know this person, but you might want to nevertheless trust their music reviews. Unfortunately, many of these essays were meant to stand on their own and start with general ranting about reputation on social networks before diving into any particular aspect. Repetition? You bet. Still, if you're willing to skim the boring parts (and if you're reading a collection of papers, you surely are), there's some good stuff in here.

Academics in a field try to figure out the most reputable folks by counting how many citations their papers get. But that's not such a great measure; some folks estimate that perhaps 20% of those citations result from someone who actually read the cited paper. If there's some accepted fact that's been around long enough to make it into a freshman textbook, no self-respecting academic's going to cite "Introduction to Something-ology." Instead, they'll look at other papers' citations to see what other people cited when referencing that fact.

An interesting study that suggests that popular music is popular because it's popular:

Research carried out by Salganik, Dodds, and Watts (2006) is enlightening in this regard. They created a music website at which users could rate and download songs by unknown bands. The home page also showed the top ten list, that is, the most popular songs… They assigned the users to eight different copies of the site with the same songs in them, without the users knowing it. The striking result was that in each of the site copies, the top ten lists determined by user ratings were different. Popularity was not primarily induced by some intrinsic quality of the songs, but by aggregated ratings activity and how it was displayed in the top ten list.

I learned of the existence of the Deliberatorium, which was a UI for finding points of agreement and disagreement on complex topics, a multi-user collaborative argument map. Have you ever wondered what large-scale decision-making (e.g., national politics) might look like when we can harness our networks and crowd-sourcing to come up with something better than representative officials? The Deliberatorium feels like a step towards that.

Tags: book social networks

blog comments powered by Disqus