Through a sequence of steps (JOHO to satterlee to JOHO) I arived to this undated thread with conversation about digital identity. 
 
Having read through several emails it struck me that there is a big misunderstanding going on through the whole conversation - people are not talking about the same thing. The fact that I pointed out several times before, is that Digital Identity is in fact a buzzword concept that has not been well defined. And what seems to be muddling in the discussion is the fact that people generally don't distinguish what shade and variation of digital ID they are talking about and for which purpose it is to be used. Some even seem to act as if a single digital ID managed by an ultimate all-singing-all-dancing digital id system was something to which the world is progressing with historical necessity.

This is a false notion. Single ID is IMHO possible, but only in a system with well defined (ideally a single) purpose. VISA or PayPal for example. Such a system does a single thing, and its owner, its infrastructure provider and users have well defined roles, and risks associated with the use of the system are well understood and managed. But when you expand the scope of the system to cover wide range of human activities, such as online shopping, bidding, inter-company contracts, education, healthcare, weblogs, online communities, MUDS ... (which all-singing all-dancing digital ID system seems to claim to do... one day) you will find out that "identity" is not a single thing.

So let me make it clear. To me "Identity" as it is mostly used in DI debate is largely artifical construct which, when stripped of its contextual attributes has very limited use. (I am exagerating little bit here. "very limited" means probably much more than what we can do at present)

Andre Durand's paper on tiers of identity acknowledges this, but I think there are few additional subtleties to what he describes.

What identity actually is (without going into philosophical issues of what "is" means) in the real world and how it is used is largely depends on context. If you read some hardcore scientific papers dealing with identity issues from legal or technical standpoint, you will find out that a simple concept of single "identity" breaks down into incredible complexities. In each context a different definition of identity with different usage mechanisms is used and simple metaphores that are used in discussions on identity such as user identification and authentication, reputation, that computer science provided us with, just can't cope with these nuances. They are not up to the task. The reason here is semantics - too many different meanings of identity, which off the shelf metaphors can't cope with.

There are other reasons beyond difference in semantics why this is not going to happen. Even when talking identity in the enterrprise context, mapping identities and getting them work within a single (centralised or distributed) system is very difficult task. The main reason for this is internal politics and costs that make such an endeavor in its full extent impractical. And we are talking about single organisation here with a single set of goals, well defined business operations and a single governance structure.

If it is difficult to achieve single digital identity within a single enterprise, what chances are it will be feasible in the context of the enterprise called the world?

9:56:09 PM