Hi there PREMIS Implementors,
I'm Brian Balsamo, the Digital Accessions Specialist in Special Collections at the University of Chicago library. Recently I've written up the beginnings of a PREMISv3 python library we intend to use for applications in our digital repository. In writing this library I've stumbled across what I believe are several differences between the data dictionary for PREMISv3 and the .xsd XML validation file that was recently published.
Discrepancies between the data dictionary and the provided .xls that I've run into are as follows:
object/objectCharacteristics/creatingApplication mandates no required child elements in the data dictionary. The “At least one of x subunits must be present if this container is included [...]” language does not appear in the Usage Notes for it nor its child elements. It is not valid if empty according to the .xsd.
object/storage mandates no required child elements in the data dictionary. The “At least one of x subunits must be present if this container is included [...]” language does not appear in the Usage Notes for it nor its child elements. It is not valid if empty according to the .xsd.
object/signatureInformation is described ambiguously in the data dictionary (“Either signature or signatureInformationExtension may be used”). Neither of its child elements obligations specify anything other than “optional” when applicable. It is not valid when empty according to the .xsd.
object/signatureInformation/signature is not repeatable according to the .xsd, despite being specified as such in the data dictionary.
In regards to the first two points:
The data dictionary does specify in regards to extensions that
"If any semantic unit is not used it should be omitted, rather
than an empty schema element included" (Data
Dictionary, page 28). Does this also apply in the case of
non-extension elements? I could find no definitive answer.
If the answer to the above is that empty non-extension elements should be omitted would it be the case that signatureInformation should specify an "at least one of signature or signatureInformationExtension" clause in its Usage Notes? This relation would be implicit if that were the case. This would address point 3.
Have I missed anything in identifying these? Assuming I haven't, is this the correct place to ask for clarification on these issues?
Thanks very much,
Digital Accessions Specialist
Special Collections Research Center
University of Chicago Library
[log in to unmask]