I'm working in the planning stages of an online publication that is
destined to combine archival documents that will be transcribed and
tagged as born-digital TEI-XML, plus digitized published volumes that
contain many document transcriptions embedded within editorial narrative
along with many partial quotations from document transcriptions. Thus
there will be at least three sorts of documentary "objects":

  1. Born-digital TEI document instances (<TEI>)
  2. Entire embedded texts (<floatingText>)
  3. Quoted portions of texts (usually as <cit><quote></quote></cit>)

Bibliographic metadata for each of these documentary objects will need
to be structurally identical or at least homologous so that they can all
be searched and retrieved using the same criteria.

Of course <floatingText> and <cit> can't take a TEI header, and the
bibliographic information within <cit> texts is often incomplete (and
may not be fully accessible within the transcribed publication in any
case). So my instinct is that the most economical way to handle metadata
will be to create an entirely separate, stand-off file (whether in TEI
or some other syntax) with a single metadata record per document, and
then simply use @corresp on all TEI elements that need to point into the
metadata records (including the born-digital documents, which would have
minimal TEI headers in that case).

Does this stand-off approach make sense to the metadata gurus out there?

David S.

David Sewell, Editorial and Technical Manager
ROTUNDA, The University of Virginia Press
PO Box 801079, Charlottesville, VA 22904-4318 USA
Courier: 310 Old Ivy Way, Suite 302, Charlottesville VA 22903
Email: [log in to unmask]   Tel: +1 434 924 9973