There is one good security argument for this situation, it's hard to trust just any old identity provider when the entry bar is very low as it technically is for OpenID. This is fairly easy to overcome via simple cooperation between the major OpenID providers. There is however obviously a business driver to maintaining control over identities, the associated profile data has value.

Facebook is notable here because earlier this year it took the step to act as a relaying party for other OpenID providers. This case is particularly interesting because the nature of Facebook means that most profile data is automatically generated rather than being teased from users in exchange for service use, so the usual downside of not getting profile data as part of being a relaying party were not that important and there was upside in community appreciation and perhaps a few more accounts. Lets hope the other sites can follow by finding their own ways of creating business value from adding OpenID relaying party support.

There is no doubt I am biased here, but when I hear people characterize XML as just "angle brackets" (or compare JSON with XML for that matter), I know they missed the point. For me it's never been about the syntax but about the the joining of two parts of the data processing world so that we could at last seamlessly mix our human languages and those of our machines in an inclusive model that did not discriminate. The implementation of this ideal in XML is far from perfect, but it meets the good enough rule for the vision and that is I think about all we can ask for something so ambitious.

So where to next? I mention services above because they represent phase 2 or what I think of as at least a 3 phase transition in how we think about data. If the merging of semi-structured and structured data with XML was the first phase, then the integration of application data silos via service & resource architectures with XML has to be the second phase and that feels like it is now also nearing completion at least in vision if not in implementation.

My prediction for the third phase has to be the opening of that data to unexpected and ad-hoc use. While I so mean this in the sense of Web 2.0 mash-ups and social networking they are just the surface artifacts of something much bigger. I think we can all feel that, but as yet don't know fully what lurks below the water of this particular iceberg.

If I was a betting man maybe I would put some money on the data portability project having the strongest vision here although it is rather too biased to the social and idealistic at times for my tastes. The software stack they promote is certainly something to note, OpenID, OAuth, RSS, OPML, microformats, RDF, apml, & XMPP, quite a collection if ever I saw one.