Eric, myself and Doc Searls are speaking tomorrow afternoon (Tuesday, March 22, 2004) here at PC Forum 2004 on the subject of social networks and federated identity.
To be honest, I can’t quite recall how this particular subject was selected, because personally, I see neither a hard nor immediate intersection of the two anytime soon. That said, as I’ve spent the past two years thinking of and writing about identity and federation, I thought I’d focus instead on some thoughts surrounding the current social networking attempts…
In early observation and through some tinkering around, in my opinion, I’m not yet sold on the current approach taken by many of the social networking sites. The reasons being…
Social Networking Systems strive to capture a social map of our connections and relationships manually rather than deduce or construct them through automatic means. Data-Collection via Manual Means = BAD & IS INHERENTLY INACURATE & ALMOST IMMEDIATELY OUTDATED.
Furthermore, networking is only a conscious or purposeful activity for a small percentage of the population.
To be a truly powerful technology, as well as both more relevent to the larger population as well as inherently more accurate, social networking systems should focus on an automated means of data-collection (the data itself has no alterior motive or ego). One way to do this might be to insert (with the users permission), agents or listeners ‘in-stream’ to our existing communications systems (such as Spoke) and then focus the attention on the algorithms which accurately reconstruct a social map of our connections as well as the relative strength & context of the connection.
Illustration of Problem Associated with the Manual Approach: How many of you have received a request to ‘network’ from an individual who you don’t recognize? I’ve received several. Now, my memory’s not too good, so often times, I’m afraid that I should know the individual, and just can’t remember them. What to do? If I accept the connection, I’ve got a rogue individual in my ‘network’ that shouldn’t be there. If I decline, then I might be perceived as rude. More on the promise of future utility and the possibility that my reputation as ‘popular or useful’ might somehow be gaged by how many connections I have than for practical reasons, I accept most of these connection requests. Furthermore, how many of you have taken the time to run through your connections and ‘tag’ them with ‘friend, business acquaintance or family member?’.
My social network will never be accurately mapped in a ‘place’ or ‘destination’ as long as the process of creating the map is external to my day to day activity.
Starting from my existing personal networks as represented by my IM, email and phonebook contacts, it should be completely possible to first mine and then automate (through smart algorithms) the process of creating a social network map, and do so in a way which bypasses the inherent problems associated with manual data-collection.