Digital Library Implementation - Content Models

Content models are formal descriptions for classes of data objects in the repository. Each content model describes a set of disseminators that the member objects subscribe to and defines each of the datastreams that they include. By definition, all datasteams in objects of the same class have the same name and mimetype. They also subscribe to the same set of behavior definitions matched up to the same behavior mechanisms. All of our batch processing will be built around content models. 

The behavior definitions can provide a bridge linking content models with different kinds of datastreams. For example, we will have three different content models for image objects because we have three cases where the content files that are the datastreams for the Fedora objects are different. For most page images, we have only a bitonal TIFF image; for art and architecture images, and for page images that have figures or for certain rare documents, we have full color MrSid images, a screen-sized jpeg and a thumbnail jpeg; for some art and architecture collections we have only a thumbnail and a screen-sized image. All of them will subscribe to at least two of the same behavior definitions and will look the same externally.