OpenID & FOAF

Another question emerging from the OpenID event, which I think Scott may have mentioned in passing, but I can't recall anyone discussing in detail, is raised by Mike Ellis here: how does OpenID fit in with the "social graph" and the various specifications which deal with "personal profiles" and other aspects of my "social network", things like the Friend of a Friend (FOAF) RDF vocabulary and XFN?

While I don't claim to be in a position to give a full answer to that question, it's worth noting that the FOAF-folk have recently introduced a property foaf:openid , which is designed to express the relationship between an agent and an OpenID. A couple of points about the FOAF approach to OpenID:

• the rdfs:range of the foaf:openid property is the class foaf:Document, so the URI used as the object of a triple with a foaf:openid predicate - i.e. the OpenID URI - denotes a document, not the agent itself. However, the property is defined as - in OWL terms - an inverse functional property, meaning that anything that is the foaf:openid of something, is the foaf:openid of no more than one thing. If I find two separate triples each associating some unidentified agent with the same OpenID URI, I can conclude that they are talking about the same agent. So in effect, the OpenID URI becomes an "indirect identifier" of the agent.
• the rdfs:domain of the foaf:openid property is the class foaf:Agent, not the class foaf:Person. Since the class foaf:Agent includes not only persons but also "organisations" and "groups", this allows for the scenario in which a single OpenID URI is indeed shared by several individuals constituting a single foaf:Agent.

And for a rather nice example building on this and using FOAF and OpenID in tandem, see Dan Connolly's piece, "FOAF and OpenID: two great tastes that taste great together", where the "social graph" obtained from FOAF data is used as the basis of a "whitelist" for authorisation choices:

... you can comment on our blog if:

1. You can show ownership of a web page via the OpenID protocol.
2. That web page is related by the foaf:openid property to a foaf:Person, and
3. That foaf:Person is
1. listed as a member of the DIG group in http://dig.csail.mit.edu/data, or
2. related to a dig member by one or two foaf:knows links.

Comments

[Mikael]

"That web page is related by the foaf:openid property to a foaf:Person"

- where is this triple coming from? You need a trusted source for this.

[PeteJ]

I was quoting Dan Connolly's piece at that point ;-)

But, yes, I assume the FOAF data they are using as the basis for this is gathered from a set of trusted sources.

So I can't just spam it with some data that says the Person with the OpenID http://getopenid.com/JoeSpammer is the Person with the mailbox mailto:timbl@w3.org or whatever.