This blog can't be viewed on LiveJournal. Instead see

Social Network Portability: First Things First

23rd Aug 2007

Brad Fitzpatrick (of LiveJournal and OpenID fame) recently posted online his "thoughts on the social graph", and started a Google Group on the subject.

There's been lots of traffic on the group's mailing list given that the group is quite young. There are lots of good ideas on there about how the world would "ideally" be, with everyone owning his or her own data and publishing it in a decentralized but secure fashion. I can get behind these goals, but as I posted in the group itself, I think this sort of talk is getting a bit ahead of ourselves. Talk about things such as FOAF and XFN has been going on for some time now, but they haven't taken hold. Why not?

Brad listed some assumptions in his "paper". Here are some assumptions of my own:

  • Most of the useful social network data is currently in proprietary systems, many of which are hostile to the idea of social network portability.
  • Most of the existing social network databases data that's also duplicated in at least one other database.
  • Figuring out how to publish data in a secure and decentralized fashion is not useful until the data is available to publish.
  • A pre-requisite for moving to a decentralized, "self-owned" model for social graph data is to level the playing field so that these new systems can interop with the incumbent closed systems.

In summary: unless you can inter-operate with the existing social network sites without their direct co-operation, you've already lost.