The announcement of Google's new OpenSocial API has generated a blizzard of commentary around the blogosphere. Yesterday Mike argued that it was a smart move
on Google's part because it creates a broader web platform that will be more attractive than any one social network could be by itself. However, the various news reports
I've read suggest that OpenSocial is missing probably the most important element of a social networking site: the networking. Most people don't join Facebook because they want to use the latest Facebook widget. They join because that's where their friends are, and because it offers basic functionality like messages and photos. Widgets are just icing on the cake.
The fundamental problem facing Orkut, Friendster, LinkedIn and the other social-networking also-rans is that people don't want to sign onto a dozen different social networking sites to keep up with all their friends. They want to sign up with a single site and see updates for all their friends in one place. As long as each social networking site is a walled garden, only allowing users to connect with other users on the same site, the largest sites will have a huge advantage because people will naturally gravitate to the site most of their friends use. On the other hand, if several sites found a way to interoperate, so that Friendster users could be friends with Orkut, MySpace, and LinkedIn users, less popular sites would be at a much smaller disadvantage.
Of course, achieving that sort of data sharing is much more difficult than simply agreeing on a common architecture for third-party widgets. Privacy would be a big concern, and it would be a lot of work to find a set of data formats that can gracefully accommodate the wide variety of information handled by different social networking sites. But achieving such interoperability would be a far more significant threat to Facebook than the features Google appears to be rolling out today. LiveJournal founder and Google employee Brad Fitzpatrick wrote in August about what an open social networking platform would look like
. Let's hope he's hard at work making sure that OpenSocial 2.0 is focused on enabling the type of interoperability he describes in that essay.