Linking online personas

People complain about online privacy all the time, but we also suffer from the opposite problem: when you want to link online personas it's impossible to do in a systematic way. Why would you want to link online information? Simple: linking data about ourselves together is how we build reputation offline. If we're ever going to solve some of the social problems online, we need to be able to do the same thing.

Here's what I mean by "linking data." When you write a resume, you're essentially linking data associated with different identities, or personas, together in a single document. You make claims about your education and past work experience that can be checked in various ways--phone calls, letters, etc.

Online, of course, you can do the same thing: put together a document that says you have a level 70 character on World of Warcraft, have 1000 positive reviews on eBay, and have written dozens of book reviews on Amazon. A human could even go verify those claims. But a machine can't. And after all, if a machine can't do it, then what's the point of having online?

Jon Udell calls this kind of linking a "lifebits service" in a great blog post on this very problem.

Suppose you're a Microsoft blogger who has launched at blogs.msdn.com. You can choose to write a mostly professional blog, or a mostly personal one, or a blend of both. Or you can separate the professional from the personal by establishing separate blogs. But no matter how you slice it, there are no good answers to some vexing questions like:

How do you integrate the online persona that you developed before joining Microsoft, or the one you will develop if you leave?

or:

If you establish separate blogs for separate purposes, but wish to combine their reputation effects, how do you do that?

He goes further and talks about the value that our reputation has when we make gestures (like linking) about others.

Page ranking algorithms are numeric, not social. People who know me, and my work, value resources I cite because it's me citing them. So they assign equal value to citations that emanate from weblog.infoworld.com/udell or from jonudell.net. But ranking engines have no idea that those two sources represent a common identity, and no idea of how other identities relate to that one.

And of course, as Jon points out, this can be about digital objects as well as people.

In the end, this will come down to standards for making and sharing these claims and checking the validations behind them. The user-centric identity technologies being developed like CardSpace and OpenID are good platforms to use as starting points. I get the feeling that reputation and claim validation will be a major theme at the Internet Identity Workshop this coming week. If you're interested, come and join the discussion.