This is the title of an article just out this weekend in Network World.
It is by David Mcaplin a senior architect at Epok. Here is an exerpt.
Metadata isn’t limited to alternative identifiers. Imagine that an XRI-identified resource is a technical manual, available as a PDF or Word document and retrievable from a variety of mirrored network locations via various protocols. In a broad sense, the manual is the same document irrespective of where it is located, how it is retrieved or in what format it is represented. XRIs are ideally suited for identifying resources at this level of abstraction because the resolution process lets the consuming application choose the best network location, retrieval method and file format for its needs from the available options.
Like URIs, XRIs are composed of an authority portion and a path portion. XRI resolution converts the authority portion and the path portion of an XRI to an XML document called an XRIDescriptor. The XRIDescriptor describes the identified resource and the means by which the digital representation of the resource can be retrieved. By providing an additional level of in direction away from concrete instances of a resource, XRIs provide a permanent, unbreakable reference on which stable business relationships can be based.
There is a great diagram on the next page.