Northwestern University's Academic Technologies group has been partnered with the Fedora Development Team since phase 1 of the Fedora project. Bill Parod answered our survey questions about the work they're doing with Fedora.
The Academic Technologies group's Fedora efforts center around managing complexity. They have web sites or collections in which the objects themselves are complex, or in which the relationships between objects are complex, and the presentation of these objects is varied, but they need to provide consistent access. They chose Fedora because it seemed to them that there was no other digital asset management system that was competing. They liked the opportunities afforded by the formalisms that the Fedora object model provides based on standards, as well as the opportunity that the software gave them to exploit existing and future functionality, in particular, Fedora’s dissemination mechanism. In general, Fedora afforded them room to grow.
In the Academic Technologies group, there are three Fedora deployments. The first is a general deployment brought up for test purposes. This deployment is used to test their object models, is being used for a class, and contains a collection being used for development on a project with three other institutions. The second is specific to a project with the Chicago Historical Society, in which the Encyclopedia of Chicago is being presented via the Web. And third, Chris Karr has set up a repository for a joint project between NWU-Academic Technologies, Michigan State University, the National Archives and Records Administration, Glasgow Caledonian University, and the BBC -- Information and Archives which will make digital audio archive materials available for use by students and faculty at the partner institutions.
In general, the Academic Technologies group has written middleware layers above the Fedora core software which interact with the Fedora APIs, including java programs that do bulk ingest and behaviors for images that drive zooming clients and allow users to retrieve specific parts of an image at a specific size.
For the Encyclopedia of Chicago repository specifically, which is built on Fedora and disseminates the repository objects as a web site for delivery to users, the Fedora objects have a range of object models, depending on the type of data being represented. The metadata datastreams in the data objects are Cocoon URLs, which act as an adapter to the mySQL database where the metadata is stored. These Cocoon URLs perform database queries on the mySQL database and produce XML. Disseminations vary in type. HTML disseminations are XSLT servlets operating on datastreams from the data object. More complex objects add to this scenario more relationship data, e.g. bound volume objects have relationship data pointing to objects that are page scans; map objects have relationship data that includes layers that can be turned on and off.
Relevant Links:
The Online Encyclopedia of Chicago