I followed the discussion of METS and OAI with interest, and I
have a question and an observation.
1) If, as suggested by Jerome McDonough and in the METS
description, complex documents were to be represented by multiple METS
objects, with an 'umbrella' object containing divs with mptrs to the
subsidiary objects, how would one get from the subsidiary object back
to the umbrella object? I do not see any kind of reverse pointer in
2) The resolution to the problem of being able to retrieve
different types of metadata from a METS object relies not only on
differences in the names of different descriptive metadata protocols, but
differences in the names of all metadata protocols, be they descriptive,
administrative, provenance, or what have you. I don't see that there is
that kind of guarantee.
Library Systems Office