I’m at a small conference on social networks put on by Microsoft Research.
During the brief intros, I make a fool of myself early on, getting it over early. I say that social networks worry me because they are based on explicit declarations of relationship, and because they’re putting valuable relationships behind proprietary walls. Well, it turns out that “social network” means something different to the academics; I meant “artificial social networks” like Friendster. Much of the room must have been puzzled.
Scott Heiferman of MeetUp gave the opening talk. Excellent, but I’ve blogged it a couple of times before.
Now a panel starts.
Joi denies that social networking tools necessarily diminish social lives and/or spirits. Blogs, he says, is publishing, but IRC is “hanging out.” Changes in presence are events, and people should be able to know about those events. Social software like Friendster filter this: Who do you want to know about your presence, and at what level of detail? Cellphones give you presence, location and mobility, none of which we’ve had in computers, and that makes a big difference.
Tim O’Reilly: We’re in the early stages of building an operating system for the Internet as a platform. We need an architecture of participation. He’s excited about Microsoft Wallop because it tries to find the existing implict data about relationships. We should be creating loose confederations that allow us to query distributed personal/social info (with the proper privacy and permissioning, of course). “We need to reinvent the user control of social networks using an end-to-end architecture…” [Right on!]
Clay‘s 10-minute talk is called “The subject of this talk is not explicit.” He wants to talk about an early mistake social network software is making. Orkut made it one-click easy to make someone a friend. The number of friends went through the roof but the network no longer reflected reality. So, they added a second click: How much of a friend? I don’t need this data; Orkut needs it to create a visible and formal model of the network. But how valuable is a formal model? There’s nothing Orkut can extract from a photo of a face that’s as interesting as what we get from it in an instant. The most important information is implicit.
So, Clay says, what led Orkut to make these wrong decisions? What is Orkut thinking? 1. It thinks that what people are doing when they think about social situations is a form of computation. This is like AI’s mistake. 2. And Orkut also assumes that, when asked, people can express they rules explicitly…but that’s false. [Loved the talk. These are topics I've been writing/thinking about, and Clay puts it all so well.]
Steve Johnson says his first two books argued against the idea that the Net consists of little echo chambers. Instead, think of it as a place in which strangers interact and new things emerge. Emergence refers to Jane Jacbob’s view of cities. [I've been reading Death and Life...a fantastic book.] He’s afraid that the new social networks are “neutering” these adventurous places. And now people — Joi, for example — are talking about the software social networks overlaying real places. He’d like to use Amazon’s Search Inside facility to search inside his own library, or the libraries of people one or two degrees away. Then he talks against the echo chamber idea: The Net is an echo chamber compared to what, he asks incredulously? TV? Even if you just follow bloggers in your general universe of interests, you’re still following links out to more diverse ideas than ever before. He points out that the criticism used to be that the Net was nothing but flame wars. Now the criticism is that it’s echo chambers. But, he worries, we are creating these social network tools in order to decrease our contact with others. [Jeez, is he good!]
Q: So, is FOAF bad, Clay?
A: No, FOAF encodes links. The degree to which you have to express a full, formal relationship will inhibit its adoption.
Categories: Uncategorized dw