I read with interest the draft proposals "Expressing Simple Dublin Core
in RDF/XML" and "Expressing Qualified Dublin Core in RDF/XML." I have
the following comments:
1) Although these drafts are both excellent, I don't agree that RDF must
be a requirement for XML documents containing Dublin Core data elements.
It seems that a step has been skipped in the progression of Dublin Core
to an XML implementation. RDF should be the next evolutionary step
(after plain XML) for users who want to implement it. I may be wrong,
but can't the RDF elements be added as a transform to a basic XML
document since they only seem to describe Dublin Core and not the data
in the document? Either way, I would like to see a separate
recommendation or standard for how to construct plain Dublin Core XML
documents.
2) The Dublin Core documentation seems to lack a description for how to
create consistent, unique and legal XML element and attribute names from
the Dublin Core element sets, qualifiers and application profiles. The
recent Dublin Core Namespaces recommendation provides the following
references to RDF documents:
http://purl.org/dc/elements/1.1/
http://purl.org/dc/terms/
http://purl.org/dc/dcmitype/
Human documentation would be helpful to describe the naming conventions
being used. I gather that the core elements are now all lower case
(title, creator, publisher, etc.) and the qualifiers can be a mixed bag
(subject, SubjectScheme, LCSH, etc.). Are there any guidelines for how
to capitalize Dublin Core element names and how to handle spaces and
other illegal XML name characters. For example the Government
Application profile includes element names such as "Date ¦ Acquired",
"Date ¦ created" and "Language ¦ ISO639-2/B". If implementors devise
their own naming conventions, Dublin Core will loose much of its
strength as a method of exchanging metadata.
Thank you.
Sincerely,
Geoff Mottram
[log in to unmask]
|