I have developed and deployed the OpenID/XRI authentication module to this site. (http://www.sakimura.org/en/)

I am pretty sure that there are rough edges, but please have a try and let me know those.

Unfortunately, current server that I use is not too stable so there might be the times that you need to retry 5 min. or so later but

Talked with Drummond this morning at iiw2007 . Finally getting to know the problem space that he is tackling. I have always been very vague on what he talks on Subject-predicate etc. and graph model etc.

Identifier usually is a pointer to an object and nothing more, but what he is trying to do is to build the relationship expression into the identifier itself.

To illustrate it, he was using

* has
* has a
* is
* is a

relationship. The example that he was drawing on was something like this.

=drummond/$is/=drummond.reed
=drummond/$is$a/+person
=drummond/$has/+name+first/(someref)
=drummond/$has/+name+first//"Drummond

Sun

It is largely the matter of the RPs.

Sun announced that it will provide the employees OpenID as long as they are employed by Sun. This is one way of doing attribute assertion.

This identifier will not be reused, Sun said.

In general, the trust model is the three way relationship amond the user-OP-RP.

By the way, there was an argument that trust can be categorized in four quadrant like this:

General
|
|
|
|
Horizontal

so that OpenID developpers are assured that you are not going to be chased after for any patent that Sun has, as long as what you are doing is not threatining for OpenID.

iiw2007

The sessions started with questions to be answered during the course of iiw2007.

Did not keep note