background image
store them in the central metadata repository.The collected information is then translated
into a uniform metadata format (the elements are based on the Dublin Core Metadata
Standard and the format is XML), and automatically indexed to be searchable.
The second, in some respects different approach is still at a quite experimental stage.
Instead of building a central repository for unified metadata, this second approach lets users
directly search at the institutional level. However, to do so, the metadata at the institutional
level needs to be structured as XML documents, which have a corresponding XML-schema
for validation.The searched metadata is then aggregated, packaged and presented, using
standard DTDs (Document Type Definition). In contrast to the first model, this second
approach gets closer to the object level. One example for this approach is COVAX, an IST-
project funded by the EC under the 5
Framework Programme.
COVAX - Contemporary Culture Virtual Archives in XML
COVAX has developed a metasearch engine that distributes searching within and across
sectors. One is able to query at the institutional level against XML repositories based on
standard DTDs or query all collections or combinations of collections (Archive, library,
Museum and Electronic Text collections).
Any DTD can be used within a collection type and multiple DTDs are possible within
each collection.
Standardised repositories enhance interoperability because they are also searchable by other
non-COVAX systems. Furthermore, future harvesting tools are simplified because of the use
of the standard DTDs and because agents can access the XML data servers via the HTTP
Queries from the user interfaces are mapped to Dublin Core (Bib1 Use Attributes -Dublin
core subset) and packaged using Z39-50 XER for the metasearch engine. Using mappings
of Dublin Core to the target elements of each DTD, the metasearch engine is able to build
an XML query for each XML database.
ALM's convert source data from European library catalogues, archives, museums, and text
centre collections to XML documents based on standard XML DTDs (presently EAD,
MARC DTD from the Library of Congress, Amico-2in1, and TEI.2). When conversion is
not possible, exploitation of COVAX could include harvesting to build alternative or
supplemental Dublin Core metadata XML repositories; centralised or distributed.
Access to XML Repositories
HTTP protocol enabled access to XML repositories,
Middleware software standards for internal communications include Z39.50, (and
Multilingual user interfaces and document labels.
Metasearch Engine
Open to any DTD and collection type,
Open to new or additional XML query languages,
User search requests are converted to Z39.50-XER format using the Bib-1 Use
attributes for Dublin Core,
Metasearch engine maps Dublin Core to the related elements within each
XML DTD and composes a query for the specific XML repository containing,