One other thing did occur to me. I am thinking there are actually a couple
of
different use cases here. One is for occurrence data which then has to be matched against a reference taxonomy, including all sorts of rough stuff.
The
other is for formatting and transfer of the reference taxonomies
themselves
(I do both at different times, also you ingest both too). Does that have
any
impact on the choices / decisions to make here?
I think this is a *key* point. DwC (and DwCA) are intended as a vehicle for transfer of names data both as *attributes* of the "real" objects (e.g., occurrences), and as "real" objects unto themselves (as reference taxonomies). Obviously, we'd expect much more noise from the former than the latter, and we can imagine services within GNA to join datasets of the former type to the latter type.
I suspect that the occurrence datasets would rely much more heavily on the "verbatimScientificName" term, whereas the reference taxonomies would lean more towards the parsed sets (although "verbatimScientificName" is still a very useful concept for the reference taxonomies, if those datasets are providing name-usage instances, and wish to provide both the verbatim representation of the name, and a parsed/canonical "code-corrected" representation of the name).
Other than that I'll probably just shut up now :)
Please don't! My mind is just now starting to get around the essence of this issue. I think it's been a very helpful conversation!
Aloha, Rich