Response to the top part of your post to follow...
On Apr 9, 2008, at 7:38 AM, Williams, Stuart (HP Labs, Bristol) wrote:
> Just a question about this document that you have in mind:
> How are you identifing it? [eg. by name, say in a filing
> system; by bit/sequence of its serialisation...]
It, in this case, is a particular sequence of bytes along with an
assertion of type. It can therefore exist in many different places,
on different hard disks and optical media, for instance, and still be
it.
I can give it many names. My file system might give it a name. I
could decide to mint a URI to name it.
(I've deliberately not used the words "define" and "identify" because
I've been trying to get clear on how to more precisely use these
words (Pat is good influence on this) but briefly, in my current
thinking, I don't think it can be defined, but it can be identified
in a number of ways, one way being by giving its name, assuming I've
given it a globally unique identifier)
-Alan