What is the federated social web?

On Sunday, July 18, 2010, StatusNet developers will be part of a federated social web summit. We'll work with other developers from other companies and Open Source projects to hammer out which technologies to use to let people on different social networks connect with each other as friends and colleagues. I'd like to take the opportunity to describe exactly what we mean by federated social web.

We're actually avoiding the term open social web. “Open” can mean a lot of different things. A federated social web is specific: it means that there are distinct entities that control parts of the system, but those parts are connected with agreed-upon rules to make a pleasing and usable whole.

Many Internet systems work this way. The World Wide Web is a federated document database. Documents are stored on different servers, but we get a single, clean interface through our browsers. Email is a federated messaging system; we send mail from one company to another without a second thought, because the federation is so smooth and painless (at least for end users).

The great thing about federated systems is that anyone can play. Any individual, company, or organization can own a Web site and be part of the Web. Any person or company can own and run their own email server.

Also, implementation details are hidden, and under your control. Nobody else needs to care how your email server works. Your implementation choices for how to run your email (Gmail? Company email system? Your own vanity site? ISP's email?) are vast. Prices for email are rock-bottom because there's so much competition.

There are other great things about federated systems. They're incredibly robust. They encourage technical innovation. They're more secure.

But our current social web technologies don't work like this at all. From the point of view of a typical social web site, if you don't have an account on that site, you don't exist. The only way for your friends on that site to interact with you is if they invite you to join the site. Despite the fact that there are hundreds of other social networking sites on the Web, almost every single one works as if there were zero other social networks on the Web.

As the medium becomes more prevalent in business and government, however, this model is changing. Many businesses are now running social networks for their employees; it's an efficient way to share information and connect to people inside the organization. Those workers need to connect to people outside the organization, too. Your employer won't want to host social networking for your vendors' employees or your customers; it'd be better if they had their own social network that connected with yours.

There's a temptation to view the momentum towards social web federation as marginal. Initial users of the Web, and of Internet email, would probably qualify as marginal, too. But Internet email and the World Wide Web have created immense wealth and social well-being for the people who used them – vastly more than the monolithic, isolated networks that preceded them.

On the other hand, it's dangerous to believe that superior architectures will inevitably win out. Making a federated social web that is ubiquitous, fun, and useful is going to be a challenge. But I think that the building blocks are in place, or almost in place, and that we're moving from the design and implementation phase to real use on the public Web. Tomorrow I'll list out the components of a federated social web, as I see them, and give an idea of where we are with the protocols needed to support them.