[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