2009/4/1 Donat Agosti agosti@amnh.org:
What I do not understand in this discussion about LSIDs, is why, for example, GBIF (e.g through the global name architecture) or ICZN both operations that would like to, or in fact use LSIDs provide a resolving mechanism for LSID as one of their first steps. It seems to me only too logic that, if you want to promote LSIDs you make sure that there is an LSID resolving mechanism around. The two are widely backed in our communities, and both depend really on LSID since they want or already go down that road. At the end, I think we need to aim at something like NIH support, similar to Genbank, that we might get if we actually are running such a service that people USE - we as scientists as well as the publishers by supporting LSID like they do DOIs.
Has anybody an idea, of how much is would cost, both in human resources and money to build and maintain an LSID registry and resolver?
If the registry only deals with authorities and namespaces it shouldn't be that complex to maintain, as shown with DOI's where the individual organisation is the only one who processes the identifying portion, and the registry is left with the very small job of taking requests and looking them up on a table and sending an intermediate redirect response back.
If LSID's are going to work the registry can't be dealing with formal registries of actual data items that fit best with the authority/namespace entity, especially if you want to decentralise the system to ensure redundancy.
Cheers,
Peter