Webwide Reputation System seen as Killer App - but is it?
Techcrunch is holding an online survey to prepare for the upcoming Future of Web Apps conference in Miami.
If you could gather together some of the smartest Web developers and ask them to brainstorm a killer app for you, what would you ask them to build? Oh, and they will only have 45 minutes to do it.Among the first roughly 1800 votes cast so far, a clear (relative) majority of readers was interested in a "webwide reputation system".
I also think managing your online reputation is one of the major challenges at the moment, and I recommend anybody interested in this to read Daniel Solove's book on "the future of reputation" and other literature around this. But I am not sure it can be addressed by hacking together some PHP scripts in 45 minutes. In fact, I am not sure the reputation problem can be "solved" like this at all. Reputation is much more complex to model than e.g. identity, which already has driven furious debates among developers, architects, users and privacy advocates.
The idea behind a webwide reputation system seems to be like this: "Wouldn’t it be cool to take my high reputation I earned on eBay and use it for Amazon? Or transfer my Slashdot karma to MySpace?" But you quickly figure that while some of these social networking and IdM platforms already have APIs, there is no real standard for interchanging reputation. In the end, it is because your reputation on MySpace does say as much about your reliability as an eBay seller as my reputation among the hacker community can convince my banker to raise my credit line. It's not a technology problem, but one of semantics and context-sensitivity.
The problem is similar to the "social graph" idea. My friends and social relations on Myspace are different than my professional contacts on XING or LinkedIn, and they are for a good reason. And so is my reputation in these different spheres, because reputation is also a relationship property. You don't have a reputation on your own, but only as a member of a more or less defined group of others. As you behave and move in different social groups and contexts, your reputations are very different across them. It does not make much sense to link these, I think. Insofar, the term "killer app" might be right: It would kill all social differentiation.
Update: Daniel Solove's book "The Future of Reputation" is now available online for free.