July 4, 2011

You join a new social web site. You invite all your friends. You rejoice that all your friends are here. You multicast some information. You lament that some of your friends don't share enough while others share way too much. Another site comes along and your friends slowly diffuse away.

When I joined Google+ last week, I was quickly disappointed that it followed the same model as every social networking site I've ever been on, going back to SixDegrees.com in 1997. Every one of these sites makes the same fatal flaw: the owners assume that this is the only web site you and your friends will ever use. This goes against the spirit of the Internet. When one router goes down, you can send your data through another. If your e-mail provider doesn't meet your needs, you can switch to another or even start your own. When one news web site gets overloaded, you can visit another one.

To me, social networking is the next step after e-mail and mass syndication. E-mail's major benefit is that anyone running a server that can send it is allowed to send it. Anyone running a server that can receive e-mail is allowed, but not obligated, to receive it. Design decisions made in the 1970s define no security, encryption, or even acknowledgement of delivery for e-mail; you can add those to a message, but there's no requirement that anyone must use them. Syndicating content over the web started in the 1990s with standards such as RSS that are also simple by design: web sites like this one expose a feed so that you can read articles using various programs and services.

E-mail and syndication share a common benefit: I can use them in any way I want without getting permission from the content owner. I can transact any business by e-mail (using encryption, ideally) and I can read five newspapers or 5,000 by plugging the feeds into a reader. Even Twitter and Facebook originally exposed RSS feeds of friends' updates, meaning that I could read through interleaved streams from all these services. Both services have since given up on interoperability in favor of forcing more users to use first-party apps or web sites. My phone has four social network applications installed, each of which works with one and only one service.

Depending on one service to express a person's identity is also a very dangerous idea. A colleague of mine, involved in the Arab Spring movements that have been all-too-closely linked to these closed social networks, had his Facebook account revoked early this year. With one action and no explanation it was like he had never been a Facebook user in the first place. A political movement should be its own social network that can be linked to, but not controlled by, a mainstream network like Facebook. That idea runs counter to the larger networks' business model of sticking as many eyeballs as possible on one web site. Any users, whether human or automated, can have their access revoked if their actions run counter to a system administrators' expectations. If you want to start a revolution, don't centralize it on one service outside your control.

This colleague became very keen on the Diaspora Project, a set of technologies that will let people and groups set up their own servers and maintain control of their own data. I'm a big fan of this as well. I see the future of socialization moving away from self-contained sites such as Facebook and Google+ and towards a more distributed model. It wasn't long ago when on-line content moved away from centralized networks like AOL and CompuServe and toward a model where anyone could set up their own web site. Some web sites became more popular than others, of course, but the cost of starting one's own site has never been lower than it is today.

I foresee a future where I can use any photo sharing site, text publishing service, video sharing service, and commerce software package, relying on a federated ID system that might include, but not wholly depend on, the big players in social networking today. Just as AOL is remembered today for bringing a legion of non-geek users onto the Internet a generation ago, so too will Facebook be remembered for introducing the concept of socialization to the open web.

The open model of socialization won't make money on its own. That's OK. It will instead let companies that sell goods and services allow customers, acting on their own, to selectively release personal data and make transactions flow more freely. I should be able to carry around a list of authorization keys that signify my relationships with other people and entities, encrypted in such a way that only I may use them for specific purposes. For example, my friend could grant me access to his Amazon wish list by generating an access token and sending it to me securely. Of course there will be a need to back these tokens up, as with all personal data, but ultimately control of my relationship tokens should rest with me. Sites like Twitter, Facebook, and Flickr already use API keys to let third-party applications use my data, but this is more to protect the larger web sites than to protect me since those sites can also revoke an application's API key at any time for any reason.

Historically, open technologies have won except where economies of scale make it cheaper for a large closed competitor to do business with consumers. The web was a win-win over closed networks like AOL, and may prevail over closed app stores for selling access to content, since content providers don't pay any royalties to a third party to serve content and customers don't have to establish a third-party billing account to access said content. Game consoles are a notable exception: closed-platform consoles sell for a loss and make up the difference on software sales, whereas open platforms like the PC require a lot more cash up front to play the same games. Despite instant messaging being a common feature of closed services, and despite the open Jabber protocol having been adopted by Google and Facebook for their IM services, mobile phones' text messaging has eclipsed Internet-based IM as the worldwide way to exchange short messages — and customers often pay hundreds of dollars a year for the privilege!

The next few years are going to be very chaotic for social networking as the market — that is, you and me — decides whether to warehouse our information on closed systems or in our own pockets. I sincerely hope that an open protocol like IMAP, SMTP, and HTTP today emerges as the standard for granting other people and companies access to selected parts of our digital lives.

Disclaimer: I work for Amazon.com. These statements represent only my own opinion and do not reflect the opinions of my employer.