[OAI-implementers] Draft of 'Expressing Dublin Core metadata in XML'

[Pete Johnston]

Members of the list may be interested to note that DCMI recently
announced the availability of a document "Expressing Dublin Core
metadata in XML" [1]. This is a working draft which has been circulated
for public comment.

The primary motivation for this work is to provide an XML format which
is based on the DCMI Abstract Model (DCAM) [2], and which provides an
XML serialisation for all of the constructs of the DCAM (or at least for
some well-defined subset). This XML format is referred to in the
document as DC-XML.

The DC-XML format is a different XML format from that specified by the
current DCMI recommendation for expressing DC metadata in XML [3]. That
recommendation pre-dated the developmemt of the DCAM, and was based on
models for a "simple DC metadata record" and a "qualified DC metadata
record".

Further details are provided in a message to the dc-architecture mailing
list [4]

For OAI-PMH implementers, the following points may be worth noting:

(i) The publication of this draft does not in any way change the nature
or status of the oai_dc XML format defined by the OAI-PMH specification.


(ii) The oai_dc XML format is an implementation of the "simple DC
metadata record" model and of the guidelines for XML encoding described
in [3]. Applications based on that XML format and on that "simple DC
metadata record" model will continue to function.

(iii) It will be possible to specify an algorithmic transformation
between the oai_dc XML format and the DC-XML format (e.g. in the form of
an XSLT stylesheet), so that oai_dc instances can be transformed into
DC-XML instances and made accessible to applications based on DC-XML and
the DCAM.

(iv) Some OAI-PMH implementers have developed XML formats based on the
"qualified DC metadata record" model and on the guidelines for XML
encoding in [3]. Again, applications based on such XML formats and on
that "qualified DC metadata record" model will continue to function. 

(v) Unfortunately, the "qualified DC metadata record" model does not
correspond to a subset of the DCAM. There are constructs in the
"qualified DC metadata record" model that do not have a clear
correspondence in the DCAM.  It is therefore probably not possible to
define a mapping between the constructs of that "qualified DC metadata
record" model and the DCAM. (Note: this was the case before this recent
draft was published).

(vi) Although a general mapping is probably not possible, the developers
of individual metadata formats based on the "qualified DC metadata
record" model _may_ be able to specify a mapping between the way the
constructs of the "qualified DC record model" have been used in the
context of their formats and the DCAM. 

(vii) If such a mapping can be made, then it would be possible to
specify algorithmic transformations (e.g. XSLT) between each of those
XML formats and the DC-XML format, so that instances could be
transformed into DC-XML instances and made accessible to applications
based on DC-XML and on the DCAM, if required. However it seems likely
that such transformations would have to be developed on a
format-by-format basis.

Cheers

Pete

[1] http://dublincore.org/documents/2006/05/29/dc-xml/
[2] http://dublincore.org/documents/abstract-model/
[3] http://dublincore.org/documents/dc-xml-guidelines/
[4]
http://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0605&L=dc-architecture&
P=4216

---
Pete Johnston
Technical Researcher, Eduserv Foundation
Web: http://www.eduserv.org.uk/foundation/
Email: pete.johnston at eduserv.org.uk
Tel: +44 (0)1225 474323