OAIS Reference Model Revisions

Proposed revisions to the OAIS reference model have now been released in draft form and are available at http://cwe.ccsds.org/moims/docs/MOIMS-DAI/Draft%20Documents/OAIS-candidate-V2-markup.pdf.  A lot to digest in here, but the more substantive changes seem to include access rights information as a new and separate class of information to be supported in an information package,  a heightened emphasis on notions of authenticity and what provenance information needs to be maintained to support that, and further explication of the relationship between representation information and the designated community, including a recognition that the necessary amounts of representation information needed by a designated community might change over time (as designated communities do not necessarily have a completely static knowledge base).  Representation Information has also now been opened up to include subtypes of information other than semantic and syntactic; this seems to be to allow people who want to include software needed to work with data in the representation information category, which strikes me as a singularly poor idea and one likely to lead to loss of information in the long-run.  The original OAIS spoke critically of relying on software as representation information for good reason (and interestingly, those criticisms appear to be maintained in the new version).  There is also a move towards allowing representation information to be maintained in a distributed fashion (i.e., for it to be held by an institution other than the one holding an information package that requires it).  While financially a good move, logistically it strikes me as problematic (you better have good alert services in place to let you know when the other institution decides to de-accession certain representation information), and logically out of keeping with the idea that representation information needs to be aligned with a designated community.  That last issue isn’t *necessarily* a bad thing; in the digital library context, the whole notion of the designated community has been problematic to say the least, as it assumes a homogeneity of technical knowledge among a community of users that does not seem to me to translate well from a physical science data curation context into a multi-disciplinary digital library context.

Which I think points to the big change going on here.  A lot of the revisions to OAIS being contemplated here seem to me to represent the intrusion of the larger digital library/digital preservation community into the OAIS space.  While that might be useful for the larger digital preservation community in terms of having a standard which addresses their needs, I’m not sure it’s such a good thing for the CCSDS and the astronomical data community they’re actually trying to serve.