1. the XRI registry probably won’t let you do that.

2. if you could do that, you probably wouldn’t be able to authenticate using the XRI. This is roughly equivalent to copying the openid.* meta tags from someone else’s web page — you would be creating a new OpenID for the other user, not creating a way to impersonate them.

You’d be better off asking an XRI expert about this though. I have not read much about the implementation details.

Does it mean that I can write under someone’s name? Assuming that I’ve registered an XRI pointing to someone’s persistent identifier.

Atom: I do realise that processing would require an additional round trip — I said as much in the article. Given that it wouldn’t affect things in the general case where is not used, I don’t really agree that this is reason enough to reject it.

For instance, it requires no more round trips than the directed identity mode, and that is included in the specification.

Section 11.5.1 of Draft 12 of the OpenID 2.0 spec recommends that OPs assign a unique url fragment to an OpenID url that changes when the OpenID changes ownership.

I believe that if the OP of http://example.com/james responded with http://example.com/id/42, the RP would be required to preform discovery on http://example.com/id/42 to verify that the OP is authorized to make claims about http://example.com/id/42. This will add an extra round trip, adding more latency to the signin process. See section 11.2 for more info. I don’t believe that using url fragments as generation identifiers will require an extra round trip.

Just thought you’d be interested.

As I said above, the main questions are (1) how to avoid someone else assuming an identity on OPs that allow identifier reuse and (2) on systems that allow a user to change their human-friendly identifier, can the user maintain their identity?