Progress Report toward Independent Identity

Last September's cover story examined the Identity Metasystem, proposed by Kim Cameron and his team at Microsoft, in support of personal identities that are independent of any vendor's silo. Microsoft's inaugural member of the Identity Metasystem is an identity selector called InfoCard and is due for inclusion in Vista when that operating system arrives in 2007. (It will be back-implemented for XP as well.)

Since then, the Identity Gang has grown in number, and it has held a series of meetings and workshops where progress has been dramatic and encouraging. In a meeting at Harvard in December 2005, Paul Trevithick of Social Physics introduced Higgins, a framework for building user-centric identity-enabled services. At the Internet Identity Workshop in January 2006 in Berkeley, creators of OpenID, LID and XRI/XDI joined various pieces to create Yadis: a new and simple combined lightweight identity system. At the Mountain View IIW in May 2006, a large conference room was packed with participants from Red Hat, Higgins, Identity Commons, XRI/XDI, the IETF, LID, Novell/SUSE, VeriSign, Tucows, OpenID and other interested parties, to engage Kim Cameron and Mike Jones of Microsoft—and to talk about open-source implementations of InfoCard.

That conversation has since been formalized in a series of phone calls and a mailing list called OSIS (Open Source Identity Selector). A report on the first of the weekly OSIS conference calls began with this:

Since then, progress has been so rapid and varied (within and between different participants) that it's hard to follow exactly what's going on. When I asked Paul Trevithick to summarize it for Linux Journal readers, he wrote back:

Two additional points. First, Dr Ahn's implementation is ready in advance of Microsoft's own. (To an enthusiastic reception by Microsoft folks at the May 2006 IIW, where the system was demonstrated.) Second, I know of at least one commercial InfoCard-compatible implementation, which should be ready by the time this issue is published.

Phil Windley, author of Digital Identity (O'Reilly, 2005) and an organizer of the Internet Identity Workshops, said:

I've been impressed, all through this process, at how committed all these different development projects are to staying open toward each other, in the general directions where they might converge. For example, InfoCard and Yadis are solutions to different problems, yet there are design decisions both communities can make today that will be interoperable at some point in the future when their uses overlap.

As we know too well, being open source doesn't prevent market-halting incompatibilities and failures to interoperate. Why, other than adherence to principles of niceness, are all these projects working to keep things from breaking as they grow in converging directions?

Phil Windley says there may be a couple of subtle reasons. First, “Sometime early last year, the competing participants got to the point where they said, 'We don't have to be enemies. We can work together.'” Second:

Instead, I think we're likely to see user-centric “independent” identity in widespread use sometime in the next two years.