There is a running discussion in the blogosphere on layers of a social networks and how trust or value is tied to each layer. This seems to have started with Joshua Porter's Bokardo: Facebook and Circles of Relationships.
A common view of social nets is a set concentric circles where I trust and rely upon the people closest to me. But then this doesn't describe reality as we experience it. I trust some people and some sources in one situation and another sources in other situations. I expect The Tribune to give me straightforward reporting on events. But my mom is a good source for what's happening with friends back home. And my co-workers have a much better instinct of how that new initiative is going to impact our project. You get the picture.
To get it even better, Sarah Cooper put together a simple Flash demonstration of this idea, Collaborative Micro-filtering. Here's a shot of that.
It's context. And it is context that most traditional "social networking" tools have a difficult time describing. The context for LinkedIn is business connections. The context for Facebook had been in-school-together, but they've shifted to a different model that I don't fully know yet (but I've joined). The context for MySpace is friends and bands. The context for Friendster was "friends" and very social connections. Orkut, the Google social network tool, has a mechanism to rank friends based on how close you are ("friend levels"), trustworthiness, coolness, sexyness (yes, really), and whether you are a "fan" of the person. But I am not sure what I am supposed to do with that data.
On top of the context-sensitive nature of my networks, they have a time sensitivity as well. If we've just met, there is often only a short time horizon over which we can reconnect and develop closer ties or learn about one another's contexts. I've had people re-connect with me months or years after meeting, and it's always a little unsettling to try to figure out what the connection was (and I've done it too). Fortunately, I have a decent archive of correspondence, so if the connection was online, I probably remember. But, still, there is a much lower likelihood of re-connecting when it's been two years since we met.
[Thanks to Patrick Lambe for turning me onto this discussion. I also happen to read Bokardo, but not as frequently as Patrick.]