[Bioperl-l] identifier interface

[Matthew Pocock]

Hi.

The actual interface for represeinting the info for regenerating a 
resource may need thinking about. While things like LSID are one way, 
raw accession/namespace tuples are another. Perhaps ensembl db name, 
entity class (feature, chromosome, clone...) and numeric id. There are 
too many scheims out there that may be usefull. For example, perhaps a 
company will want to store database locations in LDAP. In this case, you 
would want to easily implement a Bio::IdentifierI implementation that 
can store and expose an LDAP path. Perhaps you want to point directly to 
files, or to URLs, or using an sql_database_urn/Table_name/primary_key. 
Before tightly coupling the interface of Bio::IdentifierI to one naming 
standard, perhaps we could take a step back and think about the range of 
use cases. Just as genbank should not be defining the complete sequence 
interface, LSID should not define the complete identifier interface.

So, what sorts of things would you potentialy need to know to re-create 
a resource? List below.

Matthew