Digital Repositories Interoperability spec approved by IMS
The latest IMS spec to reach 'final' status deals with ways to store learning objects and make them available to other people. There are quite a few specs out there that will help systems do that already, but none specifically geared to make use of the IMS Meta-Data spec. Fortunately, IMS is not in the mood to start re-inventing the wheel, and has therefore liberally borrowed from existing specs.
In the much-vaunted learning object economy, where people would create, trade and use autonomous bits of online teaching material, the repository is a pretty pivotal part. Its where you submit learning objects for storage, expose them for searching and gathering by others, and from where you deliver them upon request. Ideally, you would alert other people when your repository is updated as well, but that's the part that IMS DRI doesn't do just yet.
But with the bits that are there, the idea is that learners, teachers and learning object authors can use their favourite VLE or Portal to submit, search and get at learning objects. Provided those systems will support IMS DRI, of course. Since not supporting it might deprive VLE and Portal users of a lot of useful content, that's not likely to be a major issue.
But submitting and querying at the VLE/Portal end is the easy bit. After that, some more sophisticated choreography is needed between the VLEs and Portals on the one hand, and databases, specialised learning object repositories and digital libraries on the other, with specialised search service inbetween. The difficulty here is essentially in translating between the different search languages, the combining of results from several repositories and simply having one 'agent' that keeps track of what all the systems are doing.
Though the functionality is potentially quite complex, most of IMS DRI should be easy for programmers to implement. Especially since important parts like the protocol for searching existing digital libraries is basically the same as the one libraries already use (Z39.50). Much the same goes for the protocol used for finding things inside learning objects (XQuery).
The picture is not quite so rosy for the mechanism used to exchange messages between all the different systems. While the basic technology (SOAP with attachments) is a well known quantity, the schema that is necessary to capture all the IMS DRI concepts in a SOAP message isn't there. Just like the public beta of more than six months ago, the final release binding document says that "An appropriate message schema is to be specified. It is anticipated that this schema will instantiate and refine more general message schemas to be developed by the appropriate IMS Project Groups."
Methods to ensure message security and ways to handle cascaded messages, failures, query cancellation and incomplete queries have not been implemented either, though the need for them is established. Given IMS' choice to go with existing specs as much as possible, that is neither surprising nor a major problem. Having half a message binding is.
The full documentation will be available from the IMS Digital Repositories page.