Interoperability and metadata
ROADS is working in a context where interoperability is becoming increasingly important as a means to integrate the wide range of information services available in today's distributed environment. Users require distributed information services to interwork in terms of search, location and delivery. The particular requirements of individual users will be varied, it is likely that provision of a variety of views onto the 'information landscape' will be required.
Within ROADS we are concerned chiefly with searching and the metadata and indexing on which searching is based. The ROADS project is investigating means by which the ROADS software can be enhanced to enable integration with other search services.
Levels of interoperability
Interoperability can be addressed at several levels, those levels with which the service provider and user are are particularly concerned might include the following (we do not refer to system architecture, CORBA, middleware etc):
Which services to search?
The user needs to navigate the 'interoperable workspace' available to them. There are various activities underway that will contribute to selection of relevant databases:
- Query routing using forward knowledge
- Formulating a Collection Description standard profile
- Z39.50 Explain
Semantic interoperability
Users will be searching a variety of indexes constructed from a number of different underlying database structures. Effective searching across services requires that semantically equivalent fields in these indexes are mapped to each other. In addition semantics in the search (client) must be managed so that they match the semantics in the indexes (targets). Activity in this area includes:
- Mapping tables between metadata element sets
- Dublin Core as a common set of core elements
- Cataloguing rules to govern content of records
- Registries, both human and machine readable, defining schemes and specifying standards in use.
Enabling standards
- Z39.50
Allows indexes to be mapped to standard sets of attributes, hiding the underlying structure of the target database.
See: <URL:http://www.ukoln.ac.uk/dlis/z3950/>
- Common Indexing Protocol
Enables routing of queries to the most appropriate database via a mesh of centroids or index summaries
- Resource Description Framework
RDF aims to provide a framework for expressing machine readable metadata about resources. It is designed to enable different applications to interoperate by using a common data model. RDF uses Extensible Markup Language as the encoding syntax. As well as offering a standard means of encoding agreed semantics it will encourage referencing of agreed definitions of metadata elements and schemes for content. Development of the framework is taking place within W3C and preliminary drafts are becoming available.
Maintained by Rachel
Heery and Michael Day of
the UKOLN Metadata Group.
Last updated: 10-Mar-1998
