14.3 MIT's DSpace System Documentation
DSpace is a repository and multitier
application being developed by MIT and HP Labs to track the
intellectual output at MIT. It's based in Java and
implemented as a J2EE application residing on Unix and using
PostgreSQL as a database. The application has been carefully
documented, including detailed installation instructions, as well as
excellent overall architecture and usage documentation.
DSpace works by allowing to the establishment of major organization
divisions, which the project calls communities. Within the
communities, intellectual output is further categorized into
collections. Each unique output item gets a
Dublin Core record attached to it and is
then combined with any external material such as images into a
bundle. This bundle is then formatted as a
bitstream, and the format for the bitstream is
attached to it. With this infrastructure in place, each output is a
complete package including the metadata information associated with
it, through the addition of the Dublin Core record.
DSpace users can submit a document or other material for inclusion
with the system, and its inclusion can be reviewed and accepted or
rejected. If accepted, the material can be uploaded; information
about the material is then available for search and browsing. In
addition, when the material is loaded, it's assigned
a handle based on the CNRI Handle System for direct access to
the material.
The type of material that can be accommodated within DSpace includes
documents in all forms, books, multimedia, computer applications,
data sets, and so on. In addition to getting access to the material
through search, browsing, or directly through the handle, users can
also subscribe to a specific collection within a DSpace community and
be notified by email when a new item has been added.
|