I think the case for opaque identifiers can be argued both ways. See, e.g. discussion here [1]. Further to Hilmar's point about offering flexibility around defining labels, I also like them because, especially for new-comers, they help distinguish the concept from the label we apply to the concept and do not give preeminence to the English label. Éamonn
[1] http://ontolog.cim3.net/forum/ontology-summit/2014-02/msg00028.html
On Wed, Sep 17, 2014 at 5:06 PM, Richard Pyle deepreef@bishopmuseum.org wrote:
I still believe that a human-friendly name is very helpful. The barrier is not the standard or how it's named. The barrier is how humans interpret and implement the standard.
I agree.
Giving the class an opaque identifier (I would, of course, vote in favor of a UUID!) would probably create a barrier to progress through opacity that is greater than the barrier of confusion through mis-interpretation of an imperfect human-friendly name like "Organism".
I think this is a false conclusion. The Gene Ontology has been widely adopted, as have been other OBO ontologies, despite (or perhaps because?) opaque identifiers.
I say parenthetically "perhaps because" because having opaque identifiers actually allows you to evolve and change labels so that they make the most sense, rather than being stuck with choices made early on because now all kinds of data uses those identifiers.
-hilmar _______________________________________________ tdwg-content mailing list tdwg-content@lists.tdwg.org http://lists.tdwg.org/mailman/listinfo/tdwg-content