RE: Abstract Schemas in XML-Queries

You are correct.  The virtual nature of the XML schema is equivalent to my
need for abstract schemas in the queries.  But an important observation is
that there is a need for the ability to map multiple schemas to the same
data.

Your two cases are identical.  The actual physical format of the data on the
server is irrelevant.  I think that is an important modeling rule.

A significant need is for the client to be able to specify the schema that
is being used in the query and for the server to be able to return a
diagnostic indicating a schema mismatch.  In the best of all worlds, a
mechanism to negotiate/discover common schemas would be nice.

Thanks!

Ralph

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


> -----Original Message-----
> From: Peter Fankhauser [mailto:fankp@darmstadt.gmd.de]
> Sent: Monday, October 02, 2000 12:14 AM
> To: www-xml-query-comments@w3.org
> Cc: levan@oclc.org; w3c-xml-query-wg@w3.org
> Subject: Re: Abstract Schemas in XML-Queries
> 
> 
> Dear Ralph LeVan,
> 
> Thank you for your comment in
> http://lists.w3.org/Archives/Public/www-xml-query-comments/200
> 0Aug/0023.html
> 
> Your use case can be regarded as an instance of Usage Scenario 2.2
> in the requirements document at http://www.w3.org/TR/xmlquery-req:
> "The XML representation of data sources may be either
> physical or virtual; that is, data may be physically encoded in XML,
> or an XML representation of the data may be produced." In the scenario
> you sketch, the database of records may not be in XML, but the server
> may provide a "virtual" XML view on the database. The view will be
> represented in the form of XML Schema or a DTD, and the database
> can be queried via this view. However, the mapping between 
> the XML view
> and the underlying (non-XML) database is not in the realms of 
> XML Query.
> 
> A more special case is to perform queries against an XML database
> with a particular XML Schema (the "server Schema") through a view
> according to another XML Schema (the "client Schema"). XML Query
> can in principle form the basis for specifying the mapping between
> the "server Schema" and the "client Schema". But XML Query 1.0 will
> not define a view definition language. A view definition language may
> be considered in future versions of XML Query.
> 
> Peter Fankhauser
>

Received on Monday, 2 October 2000 09:53:22 UTC