Like all urns LSIDs do include the "Namespace Identifier" that identifies them as LSIDs: URN:LSID:<Authority>:<Namespace>:<ObjectID>[:<Version>]
The problem is that most or even all rdf frameworks have no clue how to resolve anything else than http, hence the need for proxies. Given that, is there still the need for having LSID or even URN specific terms? Isnt it good enough to use some sameAs assertion on the resolved object if noone can deal with a urn as a link anyway?
Markus
On Oct 6, 2010, at 12:00, Gregor Hagedorn wrote:
All URNs tell you what they are by their urn prefix (not necessarily how to resolve them)
LSID do not tell this, you only know they are URNs, which means the semantic web stops here. I believe and may be wrong, that switching between a http-proxied-form-of-an-LSID and a pure LSID requires rules that semantic web processors cannot process as sameAs.
I believe this implies that if you want LSIDs in the semantic web, you need to inform about the http-proxied versions and pure LSIDs in *parallel*.
Keeping it parallel requires some design pattern. This could be a structure inside scientific name of course, but since both identifier forms are attributes of the class Pete's proposal made a lot of sense to me FOR AN RDF implementation.
Gregor
_______________________________________________ tdwg-content mailing list tdwg-content@lists.tdwg.org http://lists.tdwg.org/mailman/listinfo/tdwg-content