« DC-2006 Special session - ePrints Application Profile | Main | DC-2006 Eprints Special Session slides »

September 25, 2006

DC-2006 Basic Syntax tutorial

I'm giving a basic syntax tutorial on the second day of DC-2006 (I've put the slides on Flickr, just in case anyone is interested in what I'm going to be saying).  It's more or less the same tutorial that I have given at the last 3 or 4 DC conferences - though, of course, the content evolves slightly each year as DCMI Recommendations get updated (a process that sometimes feels like wading thru treacle - so I normally don't have to make too many changes to my slides from one year to the next).

My DC tutorials tend to follow a set pattern - I start with an overview of the DCMI Abstract Model, followed by details about how to encode DC in HTML and XML, finally touching on issues around expressing DC using RDF.  I also try to keep some slides in reserve about how DC metadata is used in the OAI-PMH and RSS 1.0, to use as case studies, just in case there's time at the end - but there never is.

This year I feel a bit frustrated, because the guidelines for encoding DC in XML and RDF are very much in a state of flux and will be the topic of discussions elsewhere in the conference.  I don't want to mislead, by giving a tutorial based on how we used to do DC! :-)  Therefore, this year, I'm going to drop any real discussion about how to represent DC using XML and RDF, which should allow me enough time to focus a little bit on the OAI-PMH and RSS case studies.

But life is never that simple!  The way in which DC is used in the OAI-PMH and RSS doesn't really fit very well with the way our thinking has now evolved.  In particular, the use of DC in RSS 1.0 is a prime example of using the dc:creator, dc:contributor and dc:publisher properties with literal values, whereas our thinking now leans towards making the range of these properties the class of all Agents - i.e. modelling the agent as a resource on which other properties can be hung.

However, making this change for the current DC properties would break the semantics of any existing RSS 1.0 metadata (and any other existing RDF metadata that uses the same convention).

Luckily, we are also thinking about replicating the 15 DCMES properties (the original DC elements) in the DCTERMS namespace.  One possible way forward, first suggested by Mikael Nilsson, is to only assign explicit ranges to the properties in the DCTERMS namespace, leaving the current DCMES elements in a state of blissful fuzziness.

The situation with DC in OAI-PMH is a little different.  Any software that processes oai_dc XML records (the DC XML record format exposed using the OAI-PMH) has to have an understanding of DC hard-coded into it anyway.  So even if the guidelines for representing DC in XML move away from our current DC XML conventions, the fact that we can make an unambiguous mapping from oai_dc to the DCMI Abstract Model means that this format will remain a useful exchange format.


TrackBack URL for this entry:

Listed below are links to weblogs that reference DC-2006 Basic Syntax tutorial:


Actually, it was Tom Baker who suggested using ranges for only the dcterms namespace...

The comments to this entry are closed.



eFoundations is powered by TypePad