Working Definitions for UVa Digital Initiatives Discussions

Resource - a logical unit of content; a resource will often be contained in a single file but may be spread across multiple files or a file may contain multiple resources.

Fedora™ - Flexible Extensible Digital Object Repository Architecture; a repository architecture that was developed at Cornell University by Carl Lagoze and Sandy Payette, which we have interpreted with our own implementation.

Digital object - an information structure defined by Fedora that associates one or more datastreams of content with a persistent identifier and a set of disseminators. Note that content is not actually stored in the digital object.

Datastream - a pointer to a resource. An internal datastream points to a resource that is in the central repository; and external datastream points to a resource that is not in the central repository. Thus far our testbed activities have all involved datastreams that point to single files, but there are interesting possibilities for datastreams to be calls to processes that return a resource. All metadata for digital objects will be stored as internal datastreams, even where the actual resources are referred to as external datastreams.

Persistent identifier - A unique code that unambiguously identifies a digital object within the central repository without respect to its machine address. Note that persistent identifiers could also be exposed to the outside world as Universal Resource Identifiers (URN) if such a scheme became widely adopted, but for the time being they will only be used internally.

Disseminator - a set of behaviors (computer programs) that an object is capable of performing. In Fedora, when an object is discovered all of the programs that one needs to use the object are attached to it. The disseminator essentially subscribes the object to an existing set of behaviors. An object can have more than one disseminator.

Central repository - the collection of all digital objects, computer programs and resources (i.e., content) pointed to by internal datastreams. Also referred to as the Central Digital Repository (CDR). The central repository will not be on one machine.

Central Digital Repository System - the system that we use to manage the digital objects and to disseminate them to our users. (Initially, the central repository system will not directly manage all of the files associated with internal datastreams in the central repository. We will have to enforce security through traditional file system controls.)

Software Application - the binary or compiled program that must be present to carry out a particular dissemination. For example, if a particular dissemination of an object requires a perl5 script to be run, the Perl5 application, or binary, must be present on the system where the process is to be carried out.