Lists Home |
Date Index |
John Cowan wrote:
> Mike Champion wrote:
> > I'm not clear on how AFs handle the problem that namespaces address.
> The idea is that there is no need to have globally unique names,
> because AFs explain how to map *your* locally unique names to
> *my* locally unique names.
I do not think that is correct. There is no way you can consider, for
instance, the HyTime meta-DTD to be a "local" name. Architectural form
elements are globally scoped by either or both of the public identifier
and system identifier. The first is supposed to be globally unique and
the second may be if you use URIs.
The definer of the derived DTD must point to the architecture. So it has
to be addressable in some sense by both parties.
> Globally unique names are now a fact of life, however; in a sense,
> the battle was won when the domainists beat the bangists.
> Part of my AF:NG effort is to allow mappings between my names
> and your names, whether or not they are globally unique.
That's fine, but it isn't the way AF:OG work.