Regarding this, I have a novel idea which I will just mention here.
It concerns URNs and identifiers, but not really metadata. So we
should move this to the URI list (whose address I have forgotten).
> On Thu, 12 Sep 1996, Titia van der Werf wrote:
> > Wonder if we should introduce sub-schemes? for example
> >
> > URL:http://string
> > <scheme> : <sub-scheme> :// <scheme-specific-part>
> >
> > or for example:
> > URN:ISSN://string
From: Jon Knight <[log in to unmask]>
> In the context of DC schemes, I don't thing sub-schemes actually buy us
> anything; we just want to be able to tell software that, "This is a URL"
> or "This is a URN" so that it can trundle off an Do The Right Thing(tm).
I do think that sub-schemes for URIs do buy us something, and
I *might* suggest that it could be generalized to a scheme hierarchy
(I'll have to think about it more). Keep in mind that this is a type
hierarchy and not a container hierarchy. It would also be a name space
hierarchy. But contrary to Jon, I don't believe sub-schemes really buy
us much regarding URNs in general. My argument on that front is already at:
http://union.ncsa.uiuc.edu/~liberte/www/URN-URC-simplification.html
which does have something to do with metadata as well.
--
Daniel LaLiberte ([log in to unmask])
National Center for Supercomputing Applications
http://union.ncsa.uiuc.edu/~liberte/
|