First, a reaction from Robin Wendler (who's having problems posting to the
list right now)
Ray Denenberg wrote:
> OBJID is described as an identifier for the "original source document".
> Maybe I don't understand what METS means by "original source document".
> But if it means either (1) the object that the METS package pertains to,
> or (2) something analogous to DC "source", then clearly this is not
> appropriate as an identifier for the METS package. No?
That is in fact the meaning of OBJID, but I'm not sure why.
The old MOAII ArchObj OBJID had the usefully vague definition "A unique
identifying string (presumably a URN) assigned to this MOA2 object".
I wonder if the definition of the METS OBJID, as we are interpreting it,
is too restrictive (or perhaps just wrong)? Shouldn't IDs for the
original source document to be in descriptive and/or administrative
metadata, rather than being an attribute of the METS element? If
people feel strongly that an identifier for the source should be an
attribute of METS, it should be a different attribute than the OBJID of
the METS object itself.
These are interesting points... It looks like Berkeley has been using
OBJID to capture some sort of shelf number for the original document,
while we, on the other hand, have been putting the (locally) unique
identifier of the METS document itself in the OBJID. So there would appear
to be a need for both kinds of identifiers in METS and the question is
which should go in the OBJID, and where to put the other.
Then there's the question of global uniqueness that is so far unaddressed.
We could, for example, define a policy of using formal public identifiers
for the unique identifiers, or else invent a new attribute for source
institution to go along with the identifier attribute.
In any case, this is clearly a great topic for the September meeting.
Robin and I can come prepared to discuss a couple of options for
identifiers of the METS objects, and those who want to capture identifiers
for the source documents can think about what functionality they might
want to drive off of that information so that we can figure out where it
MacKenzie Smith [log in to unmask]
Digital Library Program Manager phone: (617)495-3724
Office for Information Systems fax: (617)495-0491
Harvard University Library %\%\%\%\%\%\%\%\%\%\%\%\%\%