Print

Print


At 10:27 AM 7/10/2002 +0100, you wrote:
>If an implementation using METS metadata was designed both to support a
>local resource discovery service, based on qualified DC, and also to
>provide base-line Simple DC records for interoperability, say for an OAI
>repository, I imagine it will need to handle two DC formats.
>
>I guess similar situation might hold true for handling MARC or other
>metadata vocabularies?

Probably not for MARC, but that is at least in part because in
the library world, there is no such thing as 'just MARC'.  It's always
MARC+ISBD+cataloging rules+controlled vocabularies/names, and the
library community engages in an on-going process of trying to
standardize on all of these different aspects of what's in a MARC
record.  Without all those additional agreements, I suspect
we'd have a much lower opinion of MARC records' interoperability.

Qualified DC is certainly less interoperable than MARC at this point,
because we haven't done all of the work to come to agreement on
qualifiers and descriptive practices as we've done for MARC.  Simple
DC appears more interoperable, I would argue, simply because we
have such low expectations of simple DC metadata and those low
expectations constitute de facto agreement on how to use simple
DC; we've all implicitly agreed that there's no authority control on
names in DC, that you can't anticipate controlled vocabularies, you
can't rely on descriptive practices, etc.  Which means we've all agreed
that simple DC allows you to do exchange of some tagged keywords for
searching, but not a lot else.  There's no expectation with simple
DC that we should all say 'Samuel Clemens' rather than 'Mark Twain'.

What this means for METS is that endorsement of a particular
schema implementation of a metadata set, like simple DC, is
only the first and tiniest of steps towards interoperability.  True
interoperability will only be achieved by agreeing on additional
sets of practices regarding how we use those schema.  I think this was
part of what motivated the group that started trying to work on
'best practices' for METS documents.



Jerome McDonough
Digital Library Development Team Leader
Elmer Bobst Library, New York University
70 Washington Square South, 8th Floor
New York, NY 10012
[log in to unmask]
(212) 998-2425