But getting back to commenting your original post (with clearer ideas now).

Omnidirectional identity/identity projection is *very important* in that it correctly matches the way we manage identities in the real life.
- We have our national ID, medical ID, student’s ID assigned to us by different institutions;
- We control the aggregation (normally in our wallets) and disclosure of our information to others;

While today the trend seems to be towards using Open ID to aggregate your identity in one place, I argue that aggregating one’s identity in one place is bad in that it allows the creation of previously non existing connections between different closed silos – where you might be exposing different personas/facets of your identity. Leading to the possible undesirable overlap of conflicting contexts of one’s life. The aggregation of one’s various online identities should be done in a place under one’s control (from what I’ve read I’m cheering for CardSpaces :)).

I’m not dismissing Open ID. In fact, it’s the perfect match to support omnidirectional identity/identity projection!
I could have Open IDs assigned by the various entities where I’m registered (me@myCountry, me@myHealthSomething, me@myAlumni, me@myCompany) and aggregate them has part of my identity. The institutions would be only identity providers (not consumers of other institution’s IDs) and the aggregation would be made only locally/offline.

Well, using Open ID’s across sites would still be useful but only if 1 Open ID was related with only 1 persona/facet, and persona/facet management was done locally/offline.

The missing piece in getting people to take responsible control of their digital identities is this persona/facets layer (where a person would stitch together her various omnidirectional identifiers). When you start looking at identity this way you’ll probably see the web and its trends in a whole new way. At least I did. ;)

[Jon: Not sure what happened, but I found a copy of it on your blog, so here it is:]

I’ve been wondering about the same subject recently and it is my believe that somewhere along the way we dropped the chance to build a nice personas/facets layer between people and internet services. Tough luck…

I think the missing concept is that of personas. Not the limited ones same some Open ID providers support, but those described by danah boyd in her thesis http://smg.media.mit.edu/people/danah/thesis/.
Jonathan Vanasco has a good implementation of these ideas with findmeon.com (more details over at http://destructuring.net/IdentityResearch/) but I think the real solution should go even further in order to give users control over their social >unified< graph and their online presence – both managed from a persona/facet perspective.

I would like to go deeper in what I’m ranting about, but this is just from the top of my head. Eventually I’ll have to take some time to write these ideas down.

Some people perceive OpenID as a single sign on utility, without the need to correlate their behaviour at different openID consumers – in OpenID 2.0, there’s even a way to prevent that: OpenID 2 directed identity.

For others, OpenID is a way to claim ownership of a page, posting, site activity: they want to explicitly link up these disparate activities using their Openid/public identity URL.

What you call Omnidirectional identity, Simon willison calls Identity projection, see:
http://simonwillison.net/2008/Jan/7/projection/

I’ve been wondering about the same subject recently and it is my believe that somewhere along the way we dropped the chance to build a nice personas/facets layer between people and internet services. Tough luck… :|

I think the missing concept is that of personas. Not the limited ones same some Open ID providers support, but those described by danah boyd in her thesis http://smg.media.mit.edu/people/danah/thesis/.
Jonathan Vanasco has a good implementation of these ideas with findmeon.com (more details over at http://destructuring.net/IdentityResearch/) but I think the real solution should go even further in order to give users control over their social >unified< graph and their online presence – both managed from a persona/facet perspective.

I would like to go deeper in what I’m ranting about, but this is just from the top of my head. Eventually I’ll have to take some time to write these ideas down. :)

The problem is not only about finding “universal” *credentials*. It is about enabling an *identity*-conscious ecosystem to thrive (for the differences between identity and credentials see http://blogs.msdn.com/vbertocci/archive/2007/06/11/credentials-vs-identity-authentication-vs-what.aspx or http://www.amazon.com/dp/0321496841). If we consider also the omnidirectional case, that includes regulating how others access your identity rather than you just proving who you are. Proving who you are (..to whom?) may be the starting point, if you want, but what’s left to discover is what should happen before and after that in order to have a sustainable system and protect the interests of all the actors involved.
Even if the problem would be handling universal credentials, unfortunately committing to a single technology would not do. Jon mentions that PGP would not work for non-geeks, ofr example; you can pretty much find some shortcoming for every technology that will make it non eligible for some use; that’s the reason for which a meta-system is needed :-)

Yes, been there and done that. The conclusion was that these strategies were OK for me, but not for non-geek civilians.

Uh, isn’t that exactly what the PGP web of trust is? If you remain wedded to your SSL certificate, there’s there Thawte web of trust which looks a lot like the CACert assurance process – get your identity validated to earn trust points, once you have enough you can start validating other people yourself.