Dave, Eric, and Dan (posted to the list since comments on this document
are of general interest):
I finally got a chance to carefully read through your XML encoding
document [1]. Nice work and a two minor nits and questions in section
2.4.
1 In the phrase "Resources may have no, one or several identifiers and
these can be URIs. If a resource has at least one URI, the most
appropriate one should be used in the about attribute of the
rdf:Description...": followed later in the sections by "It may be that
there is no identifier for the resource, in which case neither of the
above methods need be used." What do I then do in this case? Example
(and appropriate to those who have asked "why don't you use RDF in Open
Archives Initiative"): we needed to be prudent in OAI about not
requiring a deterministic linkage from the metadata to the described
resource (item in our terminology), e.g., the publisher community was
pretty strong in saying we want to expose metadata but NOT link to the
item. So, as stated in the OAI protocol document [2] "Since many clients
of the OAI protocol may want to access the content associated with
harvested metadata, it is highly recommended that repositories use an
element in metadata records to establish a linkage between the record
(and its identifier) and the identifier (URL, URN, DOI, etc.) of the
associated item. The mandatory Dublin Core format provides the
identifier element that can be used for this purpose."
So there is a question here and something to think about. First, I
(unless I'm wront) think the prhase "neither of the above methods need
be used" should be changed to "this encoding method can't be used"? Or
is it possible to have <rdf:Description ... without an about attribute
or some other alternative. Second, if my suggestion that "this encoding
method can't be used" is true than we (the architecture committee) need
to consider the implications of that when we think about RDF as our
container mechanism. While the OAI case may sound wierd (metadata
without referencing the object) it does come from some very real
requirements.
If I'm missing some information here let me know - I'd love to see this
cleared up.
2. (extremely minor nit). The phrase "Note that the order of the
elements is not guaranteed to be preserved" is a little too abstract
IMHO. Thinking of the community who might read this document I ask
"preserved by whom"? Perhaps changing it to a more consumer view like:
"there is no requirement that applications consuming the metadata
container perserve the order of the elements in the container and
therefore metadata creators should not expect..."
Finally - I'm attaching an xml schema that was translated by xml-spy
from the dtd document. You might want to compare it to the schema you
have in the document and check the correspondence of that produced from
"hacked" (your terminology) to that produced from packaged software.
Thanks for the document and I hope these comments help.
Carl
[1]
http://www.ilrt.bristol.ac.uk/people/cmdjb/research/metadata/dc/xml/wd-d
c-xml.html
[2] http://www.openarchives.org/OAI/openarchivesprotocol.htm
---------------------------------------
Carl Lagoze, Digital Library Scientist
Department of Computer Science
Cornell University
Ithaca, NY 14853 USA
Phone: +1-607-255-6046
FAX: +1-607-255-4428
email: [log in to unmask]
WWW: http://www.cs.cornell.edu/lagoze/lagoze.html
|
