I’ve been following sporadically discussions about opening and sharing social networks’ social graphs. It’s hard to be a user of web services these days and not wish it was easier to recreate network of friends on new ones. Following sporadically means my opinion on subject may be more firm than right. I guess that makes it a perfect blogging material.

Latest article I’ve read on the subject comes from Brad Fitzpatrick, someone always worth listening to. It’s an interesting post and Brad has obviously thought about privacy issues, but there’s one conceptual problem I still don’t see resolved.

It sounds paradoxical, but I don’t think public and private data have an empty intersection. If for some reason you want to have a pseudonymous account on social network cooperating in this scheme, how can you reliably avoid being discovered? If you, for example, use same email address as with other services, then anyone using social graph data can find you out by matching hash values.

The only solution I can think of is to use a different email address, but this is neither particularly scalable and you have to decide upfront about how you intent to use the service. It also won’t work for sites, where you are already registered. There you can only hope you’ll be asked for consent before that data is given out.

Still, I’m more or less convinced that sharing will happen. Current situation is simply to painful for everyone involved (apart from biggest players) to persevere. I just don’t know what the downsides will be to which I’ll have to adapt.

Update: More on the subject from Bill de hÓra . He says exactly what I hope I would be thinking, if I was actually doing the thinking.