[Esip-preserve] Identifiers applications

Curt Tilmes Curt.Tilmes at nasa.gov
Mon Apr 12 09:43:57 EDT 2010


Ruth Duerr wrote:
> And we should also discuss the three applications we've agreed that
> need to be developed:
> 
>    1. How to register a data set and its parts (i.e., a registration app)
>    2. How users will find the ID's (i.e., a discovery app)
>    3. How users will use the ID's found (i.e., some sort of link back) 

Setting aside registration for now, it seems the other two are:

* Data Citation - A researcher gets an explicit reference to data of
  interest.

and

* Data Discovery - A researcher is trying to find some data.


I think "Data Citation" should be the focus of the identifiers work in
the Preservation cluster.

I think we should yield the "Data Discovery" job to the Federated
Search Cluster.


Federated Search is also already addressing the data registration
issue in a federated fashion rather than a totally centralized
fashion.  Their use case page discusses publishing datasets in
directories:

http://wiki.esipfed.org/wiki/index.php/FederatedSearchUseCases


I think ultimately these two will come together as "Data Discovery"
through Federated Search will lead to the Data Sets at an Archive
which will then return referenceable Identifiers suitable for "Data
Citation".

Note: Identifiers (the formal, persistent identifiers which we are
working on) as opposed to identifiers (which are simply URIs/URLs
which FROST is working with now, which may or may not have the
characteristics which make them suitable for long term referenceable
citations of data).


So, a stab at a combined Use Case:

1. Researcher wants to find some data, uses Federated Search to find
   the data -- that ultimately leads them to specific granules at an
   archive.

2. The Archive helps the researcher determine Identifiers and
   construct citations for the specific data they found.

3. Researcher does analysis and publishes the work, citing the data
   they used.

4. Someone reads the work and is able to use the citation and embedded
   identifiers to get more information (metadata/provenance/actual
   data) about the cited data.

Curt

(cc'ing the Federated Search List in addition to the Preservation
List)


More information about the Esip-preserve mailing list