Re: [Digir-dev] Question on DiGIR Darwin Core and uniqueness of Catalog Numbers

[Hannu S. <hsa...@gb...>]

>> It was suggested very early on in the formation of GBIF that 
>> provision of a service to resolve this exact issue should be a high 
>> priority.  The relative political neutrality and high profile of GBIF 
>> provides a good opportunity for a implementing a registry of 
>> institutions, preferably based on the information stored in the 
>> registry of services (data providers), that is the UDDI registry.
>>
>> The provision of unique institution codes would have provided the 
>> infrastructure necessary to support truly unique identifiers for the 
>> content being served through GBIF - that is, the specimen records.
>
> Desirable, but it is not going to be easy to make it part of 
> mainstream business...   I do not see the botanical community rushing 
> to junk Index Herbariorum and take up new GBIF assigned IDs...  Kew 
> will always be K, Leiden L, Missosuri MO and so on...
>
> But it might be possible to induce Index Herbariorum to maintain a one 
> to one map to teh GBIF IDs as part of their database so that botanists 
> gradually get used to the concepts...

Don't worry, replacing IH is not in the cards.  Rather, we are talking 
about additional infrastructure that could be used to do the following 
programmatically (via web services):

- Interprete the institution and collection codes so that data ownership 
and original provider can be recognised, as required
.
- Ensure that the codes can be resolved to something,  can be looked up 
automatically, not used wildly, and new codes can be established when 
needed.

- Find out duplicate records and various versions of the records.

- Link physical specimens to the electronic records.

- Link records to each other, including linking to other data types such 
as images and names.

- Et cetera (please add to this list)

Starting this work is now included in GBIF work programme for this year 
(items 1.2.2.4, 1.1.1.2, ...).  We are currently trying to scope it, 
hence the discussion on LSIDs etc.   The beauty of LSIDs is that the 
existing human-readable codens provided by IH and others can possibly be 
used as part of the solution. What would be new is programmatic 
resolving of the codens, for the above uses.

Regards, Hannu