KMR > E-learning frameworks > Specification development > IMS metadata > Current issues

Current issues with the LOM RDF binding

There are some important issues that need to be resolved in the current binding. First, there is an implicit set of design criteria that has been used, and these will need to be reviewed. They are, in no particular order:

Criteria
Comment
The binding should extend the Dublin Core and DCQ RDF vocabularies wherever possible. They can be found here.
The LOM 1.0 binding does this very comprehensively, but not perfectly.
The binding should reuse the vCard RDF binding from here.
The LOM 1.0 binding does this
The binding should use URIs for all vocabulary terms, and not literals.
The LOM 1.0 binding does this. In fact, I believe this is one of the most important criteria.
The binding should try to maintain the intended semantics of LOM and RDF, and thus not necessarily perfectly reflect the exact LOM information model.
The LOM 1.0 binding does this to some extent, but it is very difficult. More input would be needed.
The binding should be relatively straightforward to translate into an XML format for LOM, without losing any LOM information (other information might be lost, however).
Very important criterion, but not at all obvious. I believe the LOM 1.0 binding manages this, but I'm not sure.

The current binding has some important problems, some of which are:

Problem
Proposed solution
Currently we have succeeded quite well in the DC/DCQ compatibility, but there are several places where this is somewhat complex. These include:
  • dc:creator and dcq:created are semantically connected, as are dc:publisher and dcq:issued, but this is not reflected in DCQ. Several publising dates and/or creation dates will create difficulties, and is incompatible with LOM.
  • dc:subject is mapped to both classification and keyword
Difficult... No good DC and LOM compatible solution. Use workaround for now.
How do we model translations of titles, etc.? DCQ/RDF is not 100% clear here. Also, the current model does not support the xml:lang construct,  which we probably want, as this has recently been clarified by the RDF Core WG, and is recommended in the latest DCQ/RDF draft. Also, do we need a LangString class?
The current model has been changed.
  1. xml:lang support has been added
  2. lom:translation has been removed, as this contains information far outside what is defined in the LOM info model.
We need to update to LOM a,0 and to the latest DCQ/RDF draft. This means we are leaving IMS 1.2 behind.
Should be complete now.
What should be the recommended use of the dc:identifier element?
It is enabled. The difficult issue of what to do with catalogentry has been dealt with in a nice manner!
4.4 Technical requirements are so ugly currently. Will need to be remodeled.
Hes been remodeled, and the current implementation was found to be the best so far.
5.2 Learning Resource Type is not 100% LOM compatible.
Cannot really be solved without interpreting rdfs:range to imply Containers are allowed.
7 Relation is modeled very differently from LOM
Should probably be kept as is.
9 Classification is modeled very differently from LOM
Should probably remain similar to current implementation. Has not been changed.
Use xml:base in schemas
Done.
Typos (InteractivitType)
Done.