Jeff Jonas wrote a good piece that makes the both profound and obvious point that:

… information must be registered … or it cannot be located in any efficient manner.

He writes from the perspective of within the enterprise, and from the perspective of setting up a separate "catalog" of information. But this very idea was one of the most profound insights that led to the web: all information needs to addressable, using the "catalog" of all URLs. Once items of information are addressable, innovation can and does happen; if they are not, neither people nor machines can locate them and use them.

Which is why many of us in the OpenID community believe that digital identities must have URLs, or the people they represent won’t "exist" on the ‘net in any meaningful manner.