> > I don't by into the interpretation of
> > dc:identifier/dc:relation/dc:source to be range restricted with
> > literals.
>
> Did anyone suggest that?
Right... I was trying to argue that, for dc:relation/dc:source, the
values were _not_ literals. (I thought we were arguing that for all DC
elements a couple of months ago, but now I'm not quite sure about the
case of dc:identifier.)
> > Suppose you want to identify a human.
> > Quite often this is done by means of a (specific) passport. Why one
> > should treat passports as literals?
>
> I think having
>
> <me> <dc:identifier> <URI of my passport>
>
> to be nice. Note that the "URI of my passport" identifies *my
> passport*, not me.
OK, yes. So a "value URI" used in a dc:identifier statement is _not_ the
"resource URI" of the resource being described. It is, as you say, the
"resource URI" of the identifier-resource, a second resource.
(I'm not sure whether that "not" is "never" or "not necessarily".)
Pete
|