[Discussion of Rich et al about LSID metadata history disiderata omitted]
There is no requirement in the LSID specification---nor to my
knowledge in any TDWG standards for using it---that any history of
metadata be available. For many digital objects, and almost certainly
for "conceptual" dataless LSIDs such as might be assigned to a
specimen or an event, the most likely LSID metadata is going to be the
response to a query to a database, e.g. to return the current
DarwinCore data for a specimen record in a specimen management
database. I should think that the maintainers of most large specimen
management or observation databases would find it quite daunting to
provide the history of all changes to all the fields of their data
that happen to make up the LSID metadata, the queries for which can
themselves change over time.
Bob Morrris
