The SXSW panel was called “decentralized social networks”. The title was a bit of a misnomer — I wish there had been a panel on distributed social networks.
The presentations by danah boyd and Jonas Luster covered the well-known flaws of centralized social networks such as Friendster and Orkut — their awkwardness, explicitness, and lack of privacy.
These networks were the rage last year, but time has already proven what we guessed — without a business purpose like LinkedIn, or creative purpose like Flickr, social networks are a fad.
Tantek Celik discussed XFN, a relationship notation that really is decentralized, but bears the drawbacks of explicitness and lack of tool support.
Joyce Park had a hypothesis that the explicit definition of relationships is more appealing to men than women. Joyce speculates that women are more reliant on “little white lies”, and more hesitant to explicitly categorize their friendships. (this hypothesis doesn’t fit set of men and women I know, including men who tell white lies and women who manage social networks like collectibles).
Rather than beating up on Orkut, Friendster, and males for for being overly explicit, I’d rather see a discussion of really distributed social networks, where relationships are assembled incrementally and often implicitly.
* the patterns of conversation and interaction that are revealed by the social network analysis of collaboration through blog links, blog comments, wiki authorship, and other public trails
* comparing these patterns to the patterns in other discussion media such as mailing lists and usenet
* patterns of affiliation in creative networks such as Flickr and LastFM, where connections are draw with a combination of explicit invitation and implicite taste-sharing
Since mandatory explicitness is clearly a mismatch to realworld relationships, it would also be fascinating to:
* discuss of the social uses and time-series changes of subtler intimacy gradients in social networks like LiveJournal
* fast-forward a few years, and see whether and how the correlation of identity provided by IDCommons or a cousin has had any impact on cross-network social integration.
I must admit that I didn’t stay all the way through the panel. I was impatient for the next-generation panel — maybe next year.