On Thu, 20 Feb 2003, Phil Barker wrote:
> The direction which Scott and Pierre suggest is certainly one which I would
> endorse. A while back I tried building an Access database to illustrate
> what an relational structure for IMS metadata would look like. I quickly
> came to to the conclusion that the complexity of the table structure
> outweighed any advantages of avoiding data duplication etc.
I also agree with the thrust of recent postings.
However, I suspect that your problems implementing IMS metadata in a
relational database are a reflection of the completely unnecessary
complexity in the IMS metadata spec and its XML binding. I'm sure that
the same kinds of problems are arising for those people who are trying to
develop an RDF/XML binding for IMS/LOM metadata or who want to combine
metadata properties from both LOM and other metadata schemas like DC.
Note: it doesn't help anyone for me to say this, cos I don't expect
IMS/LOM to significantly alter the metadata spec - but I feel better
having said it! :-)
Andy
--
Distributed Systems, UKOLN, University of Bath, Bath, BA2 7AY, UK
http://www.ukoln.ac.uk/ukoln/staff/a.powell +44 1225 383933
Resource Discovery Network http://www.rdn.ac.uk/
|