Abstract Schemas in XML-Queries

A strength of the z39.50 query mechanism is abstract schemas.  The client
and server agree on a schema that they both understand and use it as the
basis for queries, even though the underlying data is not in that schema.
The server is responsible for mapping the real schema to the abstract schema
and generating the appropriate query for the physical database.

There is a similar mechanism for delivering data.  The client and server
agree to a transfer schema and the server maps the physical records to the
transfer schema.

Such mappings usually mean that structure based queries are meaningless, as
the structure of the abstract schema is difficult to map to the real schema.
But, questions can still be asked about the contents of those elements of
the abstract schema that can be mapped to the elements of the real schema.

In Use Case B.1, a DTD is defined containing an element "author".  I have a
client that is not aware of this DTD, but wants to query a database of
records built with this DTD.  The client sends in a query looking for
records with an element "dc:creator" containing the value "smith".  The fact
that the records do not contain a "dc:creator" field does not make the query
invalid, if the server understands how to map "dc:creator" to "author".

A mechanism for specifying a schema in a query is important.

Ralph

Ralph LeVan
Research Scientist
OCLC Online Computer Library Center, Inc.
http://www.oclc.org/~levan

Received on Thursday, 24 August 2000 14:05:50 UTC