Site Presented By
Saturday, Nov 7, 2009
The payoff of social networking doesn't even have to be about lust. The lure can be much more pragmatic. Tribe.net already offers targeted classifieds. What if it also offered freedom from spam?
If social networking sites are the gated communities of the Web, doesn't it make sense that when you walk through those gates and submit to their rules, you might gain the security and peace of a controlled environment?
Virtually all social networking systems offer, or plan to offer, some way of rating your friends or acquaintances. On one level this is ludicrous. If breaking the concept of friendship down into binary terms is absurd, so is breaking it into seven categories, or even 70. Such classifications may not be, as author David Weinberger has suggested, an actual "act of violence" against the concept of friendship, but they strike many people as artificial and repugnant. Pondering whether someone is merely an acquaintance, or an actual friend (if not a close friend), is not a task on which I wish to waste any of the precious remaining seconds of my life.
But such ratings may be necessary if social networks are to become "reputation systems" -- a kind of holy grail for human-relationship hackers. If I get e-mail from a stranger, but my social network rates that stranger high (because a friend of my friend rates him a close friend, for example), then I'll be more likely to let the mail through. But if somewhere along the way the sender had been judged unreliable, then the mail would be stopped.
Jennifer Golbeck is a researcher at the University of Maryland working on exactly such a reputation system -- she calls her application-in-progress TrustMail. She posited a situation in which, I, as a reporter, sent her an unsolicited e-mail. She doesn't know me from Adam, but her social network does.
"If it turns out that my advisor had read an article by you that he likes, and given you good rating ... then your mail gets a high rating," Golbeck says.
As networks get larger, they get noisier, and their utility decreases. We've all experienced this as we've moved our lives onto the Internet platform. We no longer thrill to the sound of incoming e-mail, because we know that more than likely another truckload of foul-smelling junk has just been dumped in our in box. But what if the sound of e-mail coming in rang with a chime that told us it was a good e-mail? And what if we don't even need to bother to do the ratings of our friends ourselves? Can't the network take care of that by itself? Our behavior on the Net, who we send e-mail to, what communities we join, what blogs we read, all add up to a data trail that will permit reputations to be derived automatically. We don't need to do anything but be ourselves.
Clear trends in Internet software development -- the easy merging and management of content streams via systems such as RSS aggregation, the evolution of distributed rating systems such as those employed by Slashdot, the ongoing attempt to decentralize the control of information while making it fast and easy for people and software to forge "weak ties" -- will contribute to the network's ability to determine if we are good or bad actors. This is happening all over the Net. Social software networks provide distillations of that overall evolution in tidy, user-friendly packages.
It's a classic cyberpunk construct -- humans as patterns of information, their "worth" or "value" deducible from the electronic paths they tread and the bread crumbs of data they leave behind. Yet it is hardly science fiction -- it's already becoming a fact of network-society life. So far, we're only getting our feet wet in this data pool. But we'll continue to plunge deeper.
As Bernardo Huberman, the director of the Information Dynamics Lab at Hewlett Packard Laboratories, mused to me at the end of our phone conversation, "I think we are on to something big."
