On Jun 7, 2005, at 6/7/05 11:33 PM, Renato Iannella wrote:
> Q1: Do we want to use real URN syntax?
It's the same dilemma as with XML namespaces: if it's a URI, folks
tend to assume the schema will actually appear at that network
location, and if it's a URN it begs the question of whether the
referenced document is published online and if so, where.
Seems to me like we could consider a mandatory valueList URN and a
optional valueListLocation URI.
> Q2: Do we want to use a hybrid system where we identify the
> authority that manages the vocabulary list (with a URI) and then
> indicate the value?
The atomic identity of the list is what we really care about... but
that implies a managing authority. What we definitely don't want is
multiple lists with the same "name." URIs and URNs are nice in that
they provide at least a clue as to the authority on the way to
uniqueness.
- Art
