IMS revs Meta-data spec
Wilbert Kraan, CETIS staff
June 02, 2004

But hang on, wasn't IMS Meta-Data supposed to be succeeded by the IEEE LOM standard? Indeed it is, and that is exactly what the new revision of IMS Meta-Data is mostly about: how to get to IEEE LOM 1.0 from IMS Meta-data 1.1.2. More than that, it is a pretty comprehensive , up to date and accessible guide for anyone who wants to implement the LOM.

Most specifications have an occasionally disconcerting but perfectly normal tendency to take on a life of their own after they've been released. What started out as the ARIADNE and IMS Learning Object Meta-data spec is no different: it got adopted and adapted by any number of different communities at the same time as it was further developed in IMS, and submitted to the IEEE Learning Technology Standards Committee (LTSC) for standardisation. Add the fact that the datamodel (i.e. LOM's conceptual view of the world) has been out as a standard for a while, but the XML binding (how that world view is expressed technically) is at the final stages of completion, and it is clear that there are some loose ends to tie up.

Because tying up those loose ends is the purpose of the new revision of the IMS spec, it consists of just a best practice guide and an XSLT transform with associated bits, rather than the usual IMS litany of binding, XSD, best practice and implementation guide and datamodel.

The best practice guide sums up the relevant history of the LOM, and what exactly is going to happen to it. More specifically, it outlines the subtle way in which IMS Meta-data and the IEEE LOM standard have been slightly out of synch at various points and what to about it.

The reason that the two have co-existed for a while —and probably will continue to do so for a little while— is that the IMS one always combined the datamodel with an XML binding, while the IEEE LTSC LOM working group divided the two. The thinking behind that one goes beyond just slicing up a pile of work; other technologies can be used to express the same set of metadata tags. Work on a Resource Discovery Framework (RDF) binding is coming along nicely too in the LTSC, for example.

The new IMS best practice guide goes further than setting out the relation between the various flavours of LOM, though. The bulk of the guide is a concise and readable guide to the dos and don'ts of adapting the LOM to a specific community's needs. This goes from the specifics of declaring sources for vocabulary items, to handy checklists for determining how you can determine what elements you'll need to support your community. The general lesson is simple: don't roll your own elements unless you absolutely have to.

The other part of the new revision is an eXtensible Stylesheet Language Transform (XSLT) that helps migration from the old IMS Meta-data 1.1.2 XML biding to the IEEE LOM XML binding that is currently going through ballot. XSLT is the standard technology for transforming one piece of XML into another XML format- provided you have one of many XSLT applications. if that's the case, the XSLT will do what it says on the tin.

There are a few limitations, though: it might work on the automated transformation of large amounts of metadata records, but it wasn't really designed to do that job on its own. Also, the transform presupposes that the metadata record is a separate file. That's true of ADL SCORM learning objects, but not necessarily of other IMS Content Packaging objects, where it tends to be integrated with the IMS manifest.

Finally, because the IEEE LOM XML binding is not finished yet, there is a very slim chance that something will change in the XML format that the transform is supposed to put out. That is not very likely, but it does mean that the whole IMS Meta-data 1.1.3 revision will remain a public draft until the IEEE binding is approved.


All IMS Meta-data 1.1.3 documents are available from the IMS website.