|
This page describes the Metadata Schema Repository (MDSR)
developed by the eResearch (MAENAD) group to support the
schema repository requirements of the MDE
metadata editor.
MDSR a Fedora-based repository for holding metadata schemas.
The initial version of MDSR provides some REST-based APIs
for storing and retrieving schemas based on their schema
identifier and type. These APIs are used by MDE to retrieve
schemas and by MSF to upload them.
MDSR also provides basic web-based user interfaces for
looking up and browsing schemas and for interactively
ingesting and deleting them.
MDSR uses user-defined schema keys to identify schemas. By
convention, a schema key currently consists of a schema name
and a version number; e.g. "DataCollection_v0.2". For each
schema key, MDSR can store multiple formats of the same schema,
possibly with multiple (internal) versions of each one.
Fedora also stores Dublin Core style metadata for each schema,
though there currently no MDSR APIs or web pages for searching
based on this metadata.
Implementation Technology
MDSR is implemented in Java using servlets and JSPs. It
runs under Apache Tomcat, and requires a Fedora instance
to store schemas and their metadata.
Future Work
MDSR has limitted understanding of schema formats. Schemas
are stored and returned as sequences of bytes, and limitted
validation is performed when a new version of a schema is
uploaded. MDSR needs to be made more schema format aware,
allowing (for example) semantic and keyword based schema
searching of schemas.
Other areas for possible future work include:
- enhanced support for schema metadata including an MDE-based
metadata editor, and metadata-based searching,
- improved user interfaces for browsing and displaying schemas,
- support for access control and a schema approval process,
- extensions to support thesauri and ontologies, and
- getting rid of the Fedora dependency.
Contact
For more information about this software, please contact
scrawley at itee.uq.edu.au or
chernich at itee.uq.edu.au.
|
