Rich Ziade has an interesting post over at basement.org about using OPML to define a social network. Essentially, a few prominent bloggers are concerned about the idea of vendor lock-in with your social data at social networks. Facebook is the witch being hunted currently, but I feel like MySpace is just as guilty of this - it just hasn't caught the eye of the big boys as keenly.

So essentially the idea is to decentralize your social data so that it can be easily consumed from multiple sources. This whole process is really funny to me, honestly. Imagine the past four years or so of social networking as a dialogue:

User: "You know what our problem is? Our social connections aren't clearly defined. We need one place we can go to where we can see all of our friends, and their friends! It'll be awesome!"

Developer: "Okay, we've got this great site where you can see all your friends in one place!"

User: Arrgh, one place?! That's too centralized! Vendor lock-in! Flee!"

So there's this backlash - from decentralized, to centralized, to a better decentralized, that needs to happen to be able to go anywhere on the web with your social data.

As humorous as it sounds, I think it's absolutely correct. MySpace is a great example of the lock-in - buggy code, terrible downtime issues, a closed platform. I'd wager their most valuable asset is their userbase - and its nontransferability. If all of these kids could go someplace better that they knew their friends already were, I'd bet dollars to donuts they would.

But if we're interested in decoupling ourselves from a site, and holding our information someplace else, a few problems have to be addressed:

1. How do we store it?

   There are a few ideas out there that very tangentially are trying to do this already - things like XFN (which is actually a microformat), etc. But nothing really useable yet.

   Rich Ziade comes in here with an idea of taking OPML and using it to define relationships. Essentially, each user would be an outline, and their information could be an outline within them. Interesting.

2. What's our global, unique identifier for users?

   I think that OpenID could play a very important role in this - it's already well-known as a great way to handle decentralized identity.

3. How do we keep it secure?

   This is pretty key - one of the things facebook touts as valuable is its ability to keep private things private (to the general public, anyway). So this 'social XML' would have to be kept private. So keeping it on a live site might be a bit tricky.

4. How do we get vendors to agree to this?

   This is the biggest sticking point I see. What could they possibly gain by adding this feature? They'd essentially be ensuring their impermanence. I don't see any way around this.

As great as I think this would be, I still have a very hard time believing it'll happen any time soon. Not just because the current format and technology doesn't exist yet (very rarely is this the burden when it comes to internet technologies), but because it doesn't seem financially advisable to any of the big platforms. Thoughts, nonexistent readers?