One goal of the PREMIS Implementors' Group is to determine some best
practices in implementing the PREMIS data dictionary. One means of
implementation of course is using METS as a container and including
preservation metadata as part of the METS document. It isn't entirely
clear how this should be done and there are various approaches that are
possible. Even the few documents on the PIG wiki illustrate different
approaches. The purpose of this email is to begin a discussion about what
the options are and what might be best practice-- or at least a starting
point for experimentation.
There are 5 PREMIS XML schemas, one for each entity in the PREMIS data
model and 1 which is a container. Using the container is optional. There
are several questions that need to be answered about using PREMIS with
1. Which METS section do you use for which schema(s)?
2. Do you use the PREMIS container schema?
3. Which PREMIS elements are already part of the METS schema? Should they
be recorded only in METS or redundantly?
I think the first two questions could be considered together and the last
is a separate one. So I will start a thread regarding which sections of
METS should be used.