RE: "Z39.50 Next Generation"

>      - This is an experiment, not a standards process.
>      - This is one of possibly several such experiments, to test
>        approaches for evolving Z39.50.

I think (in hindsight) giving the name ZNG to this experiment emerges was a
mistake. This is a group of implementers "playing" with a subset of Z39.50
over SOAP in the hope of reporting back to the ZIG on various directions
Z39.50 could take *after having some practical implementation experience
rather than just discussing the philosophy*. The ZIG is afterall an
*implementers* group.

In the WSDL description we're constructing, at present I'm calling it
prototype 1 not version 1 to try to avoid any confusion that this is meant
to be a final form.

This is only one such experiment - I'm personally involved in two others (I
sent brief details of these to the list a few months ago, but received very
few comments):

A SOAP Search developed by a small group from the GILS community. This is
very similar to ZNG except that it retains RPN albeit in an XML sanitised
form and retains multiple query languages and result formats albeit using
xmlns rather that OID's. We deliberately position this as being "derived
from Z39.50" rather than a new version of Z39.50 or GILS etc. (see
http://www.gils.net/search.html)

A template WSDL description of the entire Z39.50 version 3 protocol. The
intention is to generate this automatically from the ASN.1 via XER and to
having bindings to both SOAP (i.e. implement the entire v3 protocol without
change but over XML using XER over SOAP) and a binding to BER over TCP/IP
(i.e. the WSDL is a fully functional alternative to the ASN.1 description).
(see http://www.lib.ox.ac.uk/jafer/ez3950/z3950-wsdl.html).

I was hoping that things such as the SOAP service above or ZNG might be
simple subsets of the full WSDL to encourage an easy to implement protocol
which could be extended to the full Z39.50. 


Matthew Dovey
Oxford University

Received on Tuesday, 17 July 2001 05:10:04 UTC