Social networks 2: The revenge of the social network

So ok, basically what I talked about in my previous entry is related to something that is pretty much a glorified Plaxo or something like that. It’s slightly more complicated because you would need a relatively complex process for handling multiple identities and all sorts of other gray areas but by and large I think I can envision something that nobody can actually provide today. The problem is pretty much what value this actually adds? If we have a large amount of existing social networks, what would the benefit be of something like this. Well, it becomes interesting at the point where we can load it with other functionality.

For argument’s sake, let’s say I have now created a system whereby I can automagically creates social network nodes based on my contacts. What value can we load this with to ensure that my contacts actually become users? (i.e. how do we turn this into a business model?) Something that I have been interested in for quite some time in relation to social networks is what I call an interaction (or conversation) index. I look at it this way, take my good friend paze for example. He’s  a tech geek just like me, slightly different in terms of his particular interests but we have worked together in the past and have a number of overlapping areas of interest. Over the years we have had an ongoing conversation, a large part of which has been carried out online in IM or email. So right there you have a decent amount of text being created. The conversation (or conversation of conversations if you will) has covered a multitude of subjects since certain subjects have been intimately tied to a project bound in time, others pop up again and again. If we could index these conversation, find common subjects within them, rank them depending on when and how frequently they occurred it will tell us something. (There is actually a lot more to this idea but I don’t want to give everything away here) So now, not only do we have a social network spanning a large amount of users (if we have done our viral homework and the service we provide is sufficiently interesting), we also have another network that we can call an interest network. Let’s face it, even though one of the people I communicate most with is my wife (generally SMS and Skype messages) these conversation don’t really spill over into subjects that are part of my professional life. But if we could analyze my conversations with peers and colleagues we could probably find that there are a number of keywords that keep popping up and that define my relationships with different people. In the case of paze that would probably be alcohol, stupid ways of cluttering the namespace with variable names that are inappropriate and various programming languages. So now I have two maps, one of my social network and one of my interest network. I use interest in the loosest possible sense here, in the case of yours truly I work with and have worked with a lot of things that I am not the slightest bit interested in but would nevertheless pop up as subjects. Those would however be local in time generally and fade fairly quickly once they are off the horizon.

So. An interest map and a social map. We do have a trust problem though. We would need to be a trusted entity in order to do this. But hey! – people trust Myspace with a lot more than this so let’s ignore that particular problem for now. What other information could be useful? Well bookmarks would be. Why is that? Ok. Let’s say I am currently very interested in concurrent programming because of some bizarre project that I am working on. Somewhere out there is Larry Otherguy who happens to also be very interested in concurrent programming. He decides to bookmark a page that has some information about that particular subject and since we – somehow – have the ability to register what he bookmarks we can index that page, we can try to correlate it with his interest map (i.e. what keywords are relevant for this particular resource) and then see if that information can be distributed somehow. It may not be useful to me but chances are that it has a higher risk (if you will) of being relevant than any random page in a search result that we could get from <insert relevant search engine>. So we are trying to leverage the social networks a bit by indexing conversations (email and IM in this case – never mind how we do it technically, let’s allow this to be a little thought experiment for now), creating social and interest maps and leveraging the ability to spread information based on simple indexing and correlation towards the interest maps of several individuals. We could create a search engine that works with the bookmark space of the entire network only, giving us even more interesting capabilities with “free” metadata. But for now this will have to do. Look out for part 3 a little later if I come up with some other stupid ideas to bother you with.

Now I will cook some dinner and ponder what else we could do with this.


About this entry