EP: I use and like both services; I think they’re important for the web ecosystem. However, I think that the true benefits of social software – social networks like Facebook, social messaging like Twitter – are going to come when federation is easy to use and widespread. I see no reason that Facebook and Twitter can’t participate in that future federation. After all, don’t we all have friends with @aol.com email addresses?
LU&D: You were one of the pioneers of the distributed model for social sites. Did you always have in mind building towards something more than a collection of microblogging sites, in terms of functionality?
EP: I did not. It was only while talking with developers of other social networking tools – proprietary and open source – that I realised that the essence of social networking software is distributing activity updates to an opt-in group. Since that’s what StatusNet does very well, we decided to test out using it for similar ‘microcasting’ applications: social music, social photo sharing, bug tracking systems. We think we’re onto something. There are going to be some tough parts, like distributed private messaging. But I believe we’ll get there.
LU&D: Identi.ca/StatusNet has great features, like context view and groups, not seen in Twitter – why haven’t users noticed this?
EP: One of the most important principles for online communities is Metcalfe’s Law. It says that the value of a communications network to each participant goes up with the square of the number of participants. As we hear it from end users, “I use ‘X’ because all my friends are there.”
The upshot of Metcalfe’s Law for federated communications systems is threefold. First, it makes the prospect very attractive for small, isolated systems; value goes up dramatically for their users if the networks federate. Second, it makes larger systems resistant initially to federation; there’s no incentive for them to join, and the downsides (lack of control over the network, incentive for users to invite their friends to join) are higher, relatively. Third, it means that as federation picks up pace, the larger ‘network of networks’ gains momentum, and eventually even the very large networks need to join.
Consider email, for example. Early on, the drive for distributed internet email came from small ISPs, universities, corporate internets. Large consumer systems like CompuServe and AOL didn’t allow sending internet email. Only after the number of people with internet email addresses was much, much greater than the number of AOL subscribers did that company eventually open up to SMTP. I think we’ll see similar adoption curves for distributed social networking.
LU&D: How do we get users to care about who owns their data?
EP: I think that’s a really difficult sell. It’s boring and pedantic. Compared against the fun that social networking services provide, talking about privacy issues is really a downer. Who wants to worry about obscure marketing issues when there are friends-of-a-friend to send flirty private messages to?
It’s much more likely that change comes from another direction. There are entities that simply cannot accept turning over their data and online presence to a third party: governments, political parties, corporations. As these organisations become more engaged with social networking, and want to get more engaged with each other, they’ll insist on a federated approach that gives them full control of their data and presence.
This interview appeared as part of Linux User & Developer’s feature Social Networking: The good, the bad & the ugly.
Want to know what else featured in issue 90 of the UK’s leading pro-level Linux magazine? Click here.