I think we're talking about two different types of "things". Thing 1 is your car. Thing 2 is Picasso's Guernica. This is how I see XDI addressing applying to them:
I completely agree with what you said, i.e., that "My car is an actual thing. It should be discoverable. It should be able to have relationships. It should be portable. It
In other words, your car should be just like a person in XDI, i.e., it should have an immutable XDI address (cloud number), and then it could have one or more mutable XDI addresses (cloud names). Example:
[*]!:uuid:f81d4fae-7dec-11d0-a765-00a0c91e0001 <== cloud number for Les's car
=les.chasen+car*hotrod <== cloud name for Les's car
=les.chasen+car*hotrod/$ref/[*]!:uuid:f81d4fae-7dec-11d0-a765-00a0c91e0001 <== XDI mapping statement
[*]!:uuid:f81d4fae-7dec-11d0-a765-00a0c91e0001/$is+/+car <== XDI typing statement
All good. This means Les's car is a top-level XDI authority in the * (thing) space. Per our other parallel conversation about XDI authority subgraphs, a copy of the XDI authority subgraph for Les's car could sit at Les's XDI endpoint, or any other XDI
endpoint that needs data about that car (such as his mechanic's).
Important note: the authority subgraph for Les's car in the [@] space is DIFFERENT THAN AND PEER TO Les's personal authority subgraph in the [=] space. So both Les and his wife and his mechanic can share access to the same authority subgraph for his car.
This is a highly unique global icon that will always have the same human identifier (I believe, anyway), so my only question is whether it needs a cloud number. Let's be generous and assume it does. In that case it looks just like Les's car but in a different
namespace, i.e., it is identifiable in the top-level * space.
[*]!:uuid:9ce739f0-7665-11e2-bcfd-0800200c0002 <== cloud number for Picasso's Guernica
*picasso*guernica <== cloud name for Picasso's Guernica
*picasso+painting*guernica <== cloud name for Picasso's Guernica
XDI mapping statement
XDI mapping statement
[*]!:uuid:9ce739f0-7665-11e2-bcfd-0800200c0002/$is+/+painting <== XDI typing
So now you have the full parallelism between how you identify and discover Les's car and Picasso's Guernica. But whereas Les's car would be discoverable in Les's XDI graph, Picasso's Guernica could be discoverable in a community XDI dictionary that includes
entries for *names.
How's that sound?