Re: Next ZIG - new record type / How to request a spcific schema for XML records?

At 09:00 08-03-01 -0500, Robert Waldstein wrote:

>  So I think the proposal is to make OIDs a choice with a STRING everywhere
>they appear in Z39.50 -  can this be done in a reasonable time frame, and
>does it require reballoting the whole standard?  Or is this an over-
>simplification and each OID needs to be looked at?

Would there be a golden middle ground in *only* changing the schema 
identifier to support string types to begin with? In an XML world, it is 
clearly the most volatile element: We don't invent new record syntaxes that 
often, but everybody and his neighbour has their own schema.

Thes requires a modification to the Specification structure under 
CompSpec... It *IS* part of the protocol, but, I'd wager, a much less used 
part of the protocol than the Record Syntax.

The other place where I predict we will soon need to identify XML schemas 
is in the query, where we would like to address specific elements by tag 
path (preferably by using the existing XPath syntax) rather than by USE 
attributes. In a community that defines itself by XML-based data models, 
the tag-paths, after all, are probably the closest thing to abstract access 
points that you will find. However, this could be handled simply by 
introducing a new Query type - the same way it was done in Z39.50-1992 to 
support proximity queries.

Eventually, there are a number of little places in the protocol that would 
be touched by embracing an XML-based view of things (using a very broad 
interpretation of the term "XML-based"). These things should be done, but I 
think its possible and desirable to do so in an incremental way which 
allows gradual development, implementation, and experimentation with a 
minimum impact on interoperability.

--Sebastian
--
Sebastian Hammer        <quinn@indexdata.dk>            Index Data ApS
Ph.: +45 3341 0100    <http://www.indexdata.dk>    Fax: +45 3341 0101

Received on Thursday, 8 March 2001 11:10:39 UTC