RE: Next meeting

A group of us have been working on an XML query structure based on bib-1,
but I would hesitate to call it Z39.50 at this stage. Details are at
http://www.gils.net/search.html

As you can see we've moved the operations into element names, also we've
moved the attribute set definition into an xmlns attribute. We have also
moved the attribute type into an element name rather than an attribute. This
is because WSDL/SOAP recommends using elements over attributes.

I'm just about to argue to drop the binary aspect of the operations.

Matthew

-----Original Message-----
From: Denis Lynch [mailto:DLynch@ACM.org] 
Sent: 18 May 2001 06:46
To: rden@loc.gov
Cc: ZIG
Subject: RE: Next meeting

It's good to see that the Z39.50 crowd is at least aware of what's happening
in the real world :-)

I never had much hope for XER. Looking at it, that seems pretty well
justified. If it's ever going to go anywhere, it needs to be edited by some
folks who understand XML.

Some of the issues could be dealt with even at the level of mechanical
translation: simple types (especially integers and enumerated strings)
should be attributes, not elements. The Query structure looks even sillier
in XML than it does in ASN.1. Part of the reason must be that it's not even
an accurate translation, (an "op" isn't a choice between an operator and an
operand, is it?) Even if there's a desire to keep the model of binary
operations, the operations should be element names, so

<Or>
 <Op>
  <Attribute Set='Bib-1' Type='Use'>4</Attribute> . . .
<Term>Denenberg</Term>
 </Op>
 <Op> ... </Op>
</Or>

The WSDL work Matthew Dovey is doing looks quite promising!


--Denis

Received on Friday, 18 May 2001 07:16:21 UTC