More on XDI contexts and RDF blank nodes

["Drummond Reed" <drummond.reed@cordance.net>]

On today’s telecon we the meaning of XDI subcontexts relative to the XDI statement that contains them.

 

Per the V9 XDI RDF Model proposal (http://wiki.oasis-open.org/xdi/XdiRdfModel), in XDI RDF there are three options for an XDI RDF object:

 

1) Literal

2) XRI

3) XDI subcontext (i.e., another XDI document).

 

This lines up with the three options of an RDF object:

 

1) Literal

2) URI

3) Blank node.

 

Conceptually that means an XDI subcontext with an RDF blank node. However Bill pointed out today that the notion of an XDI subcontext is different than an RDF blank node because an XDI subcontext may contain an entire XDI document, i.e., multiple XDI statements about multiple XDI subjects, not just one blank node. Thus he concluded that XDI subcontexts are more like nested RDF graphs.

 

I think he’s right. However I realized that in such a subgraph, there is one XDI RDF subject -- $ -- that appears to emulate precisely the properties of an RDF blank node. In other words:

 

1) There can only be one such subject in the XDI subcontext (just like an RDF object can only be one blank node).

 

2) Without another identifier for this node, this subject cannot be identified without reference to its parent context.

 

3) Like any XDI RDF subject, it can have any number of its own predicates.

 

That means that the concept of an XDI subcontext is actually a superset of the RDF concept of a blank node. An RDF blank node is *one possible subject* in an XDI subcontext. And in XDI RDF, that particular subject is always addressable through its parent context:

 

      subject-xri/predicate-xri//$

 

 

=Drummond