Our repository uses an 'atomistic' content model, namely one in which each digital object is spread across several fedora objects. This makes each fedora object very small and flexible. Using this approach does require greater use of fedora's resource index (so hone your Sparql skills).
This is a “core” object in fedora. These objects are indexable by Solr, and they contain the following datastreams:
This is a content model for a “simple” image-based object. There are no required datastreams per se. These objects may have relationships to objects with the following content models:
This is a content model for manuscripts. These objects typically also have an amherst:objectCModel model. Like amherst:imageCModel there are no required datastreams.
This is a content model for manuscript pages. These objects typically also have an amherst:imageCModel model. There are no required datastream. Each of these objects should be related to the corresponding manuscript with an isPartOf relationship. These objects also use the http://acdc.amherst.edu/relationships# namespace to describe isPrecedingPageOf and isFollowingPageOf. Like amherst:imageCModel, there are no required datastreams.
This is a content model for the original (archival) image stored in the repository. It contains the following datastream:
This object should have an isPartOf relationship to the core digital object.
This is a content model for the VRA Core metadata for a particular object. Objects with this content model should define an isMetadataFor relationship to the core amherst:imageCModel. They contain the following datastreams:
If EXIF metadata is available, then it should be in a separate object, stored with an amherst:exifCModel model, related to the original image through an isMetadataFor and isDerivationOf relationship. The metadata itself will be available in the only required datastream:
This is a content model for a JPEG 2000 image. The JP2 image is stored in the only required datastream:
This item should also contain an isPartOf relationship to the core object and an isDerivationOf the archival image.
This is the content model for a simple JPEG image. The digital object will contain the following datastreams:
Each of these digital objects should be related to the archival image through an isDerivationOf relationship. They should be related to the core digital object through an isPartOf relationship.
This is the content model for a collection object. It will typically only have a RELS-EXT datastream.
This is the content model for a personal collection. It, too, will typically only have a RELS-EXT datastream.