I had a chance to sit down with David Glazer, a director of engineering at Google, and talk with him about OpenSocial and related projects. From this man on the inside of the project, I wanted to get a picture of what OpenSocial will mean to Web users, and find out when.
What it is, what it isn't
OpenSocial is an application platform that big Web sites and social networks can choose to support. If a developer writes a product for OpenSocial, then users on any network that supports OpenSocial will be able to run it on their profile pages. For example, the TripWiser application I saw last night is an OpenSocial product. It should run on sites, such as Orkut, Bebo, MySpace, and Hi5, which support OpenSocial. Running the application on each of those platforms will let users connect with their friends on that platform to plan trips.
What OpenSocial does not do is interconnect the social networks themselves. For that, there's a slightly newer emerging standard called the Social Graph API. Products written to this platform will be able to read and write friend data between social networks.
OpenSocial is a big deal because it makes it easier for developers to leverage their work across multiple social-networking sites. This makes the economics of developing social applications better and should lead to better applications for all of us. But the Social Graph API is a much more important project, because it lets users combine their personal networks in the ways they want, unfettered (theoretically at least) by the artificial walls that the social network sites put up around their audiences.
Glazer gave me a good example of why users will learn to love the Social Graph API: Say you're a big user on a general social site such as MySpace. Then you head over to the specialized pet-focused site, Catster. "Who's there that you know?" Glazer asks. The Social Graph API will let you find out--it will tell you which of your MySpace pals are also on Catster, helping you get up to speed very fast on the new network.
Saying it doesn't make it so
There are many tough problems to solve before everyone, and every social platform, will support OpenSocial and Social Graph. Technically, each social site is different, so writing an Open Social application doesn't guarantee a successful release of the application onto all sites. But it makes it much easier.
From a business perspective, the Social Graph API turns the world on its head. Why build a social network audience if your users can take their friends anywhere? There are reasons, of course, but opening up the walls of the gardens do change the business dynamics substantially, and the companies that want to let users move their friends in and out of networks need to rethink what it is they have that makes them valuable. As Glazer said, the answer to this problem is unknown, but it's certainly not, "hold the users hostage."
Then there's the privacy issue. Do you want your friends taking your information from one site you're on and pulling it to other services? Glazer's initial take on this is, "It's just like search." Meaning that the Social Graph API should only have access to public profile information. That's too limited, we agreed, although the other extreme is too scary for most users. It's an unresolved problem.
What we'll see
The first OpenSocial applications (not necessarily Social Graph-enabled) will start rolling out the early adopter sites such as MySpace, Orkut, and Hi5 this year. Users will just start seeing more cool applications on their favorite platforms.
But things won't get really interesting until these applications start to communicate across networks. Then a contribution that a person makes on your page on Bebo might also show up on a mutual friend's installation of the same widget on Hi5 (but not Facebook; it's still its own island). I saw how that might look last night during the demo of ReadingSocial, a reviews service that could collect reviews on topics across networks. Think of it as a decentralized Yelp. It's a powerful idea and it indicates how cross-site social networking could give power in social networking to users, by removing it from the current major platforms.
See also: Pizza time for OpenSocial applications, by Dan Farber.