1. Profile Information (user data)
2. Friends Information (social graph)
3. Activities (things that happen, News Feed type stuff)

Looks like OpenSocial won't have its own markup language (Facebook, for example, requires the use of FBML), which means developers will have one fewer thing to learn. The platform is launching with a bunch of partners in place, including Orkut (owned by Google), Salesforce, LinkedIn, Ning, Hi5, Plaxo, Friendster (apparently they still exist), Viadeo and Oracle (not well known for social networking, but they've got an impressive sounding internal app). They've also worked with some of the bigger Facebook developers to get them on the new platform, including Flixster, iLike, RockYou and Slide.

This is interesting, if for no other reason than that Google has managed to perform a buzzword bingo coup and cram two of the hottest words of 2007 into a single name. Sometimes it feels like the sky is raining social from giant buckets, and you'll be seeing a lot more about open in the coming months as the open source world continues its mainstream push. We're big believers in (and supporters of) openness and I'm glad to see some of the walls around the garden come down. The is brilliant in a lot of ways (particularly in their ability to scale and to solve the security issues which typically ruin similar efforts), but your data is very much locked into their site. We'll have a better idea of how OpenSocial will change that when the API is actually released later (and particularly about whether a layer could be developed which allows developers to build Facebook apps on it), and will report back.