CORDRA or reconciling local repositories and global federations
Wilbert Kraan, CETIS staff
March 25, 2005

These days, many elearning community virtually define themselves by setting up their own repositories, each with its own technology choices, metadata and content. Which is fine, but it can mean that the community can miss out on great resources made elsewhere. The CORDRA (Content Object Repository Discovery and Registration/Resolution Architecture) initiative seeks to reconcile these local concerns with global reach. A report of a two day workshop on the topic is now available.

Few things get the central problem of interoperability in such focus as content repository interoperability: how to square the priorities of a particular community of practice with the desire to re-use what other people have already done.

Whether the community is defined by geography, subject, educational level or any other aspect, they all have their own needs for content format, how to describe that content and much more. Crucially, the community is practically defined by who is allowed to share resources, and under what conditions. If agreement can be reached on all of these issues, the logical next step is to buy or build a repository that supports the preferences and choices, and regulates who can share what.

Such a community has a natural limit; it's exceedingly difficult to get agreement on all of these issues with the rest of the world, after all. But some people in that 'rest of the world' are bound to have a number of resources that could be of great value to your community. The logical next step, then, is to federate all these repositories.

Except that federation necessarily means that agreements need to be reached between communities before they agree to throw their treasure troves into the bigger pot. Not just that, where a repository is a reasonably well understood piece of technology, the best way of federating repositories is still a matter of some debate.

Hence the first workshop on CORDRA in Melbourne, where both the implementation of federations, and their practical and political ramifications were discussed by members of a wide variety of communities.

A single outcome is hard to distill from such a wide ranging exchange, but the crux remains who controls what at which level. From matters of access control, to staff motivation, to metadata exposure, to search or harvest protocols, each requires tough compromises between what a specific community wants, and what a cluster of communities agrees in a federation.

Most workshop participants agreed, however, that CORDRA has some strengths in the fact that it is a model that puts very few constraints on what a repository in a CORDRA federation needs to be able to do. Also, it leaves the major policy choices about the federation itself to whoever cares to join that particular CORDRA federation. In short, CORDRA as a model anticipates diversity and limits centralisation to what people are prepared to bear.


A full report on the International CORDRA Workshop is available online in D-Lib magazine.