RE: [search-ws] Untidy Arrangements: recordXMLEscaping and recordPacking

["Ray Denenberg, Library of Congress" <rden@loc.gov>]

Tony - 

I've changed recordPacking and recordXMLEscaping to optional.

I took a look at the current 1.1 spec and noticed that recordPacking is
mandatory on the response (see
http://www.loc.gov/standards/sru/sru1-1archive/xml-files/srw-types.xsd)
however I agree with you, they both should be optional, and I have changed
them both. 

 
On the question about:

<xs:any minOccurs="0" maxOccurs="unbounded" namespace="##other"
processContents="lax"/>

This satisfies the desire to be able to gratuitously include arbitrary
elements from foreign namespaces, as you can do in ATOM and RSS. We talked
about this but I can't readily find that discussion - If you are compelled
to wrap them in extraRecordData and extraResponseData then you lose
flexibility, however we don't want to drop those elements because then we
would lose compatibility with earlier versions. 


This schema is (now) intended to be valid for all existing versions - 1.1,
1.2, and 2.0. That's also a decision we made several months ago.  So a
parameter that's in 1.2 but not 2.0 is in the schema but is optional.  Even
though it may be mandatory in 2.0. So it isn't a completely tight schema, in
that sense.  But that's why resultSetIdleTime is still there. 

So version is still in the schema for the same reason. But it's value should
be 1.0, 1.1, or 1.2.  2.0 would not be a valid value. (I'm not completely
sure what your question about version was.)

--Ray

-----Original Message-----
From: Hammond, Tony [mailto:t.hammond@nature.com] 
Sent: Thursday, June 17, 2010 12:21 PM
To: Denenberg, Ray; 'OASIS SWS TC'
Subject: Re: [search-ws] Untidy Arrangements: recordXMLEscaping and
recordPacking

Hi Ray:

Thanks for this. Is looking good but was getting some errors with validating
my test with facets. Will need to spend some more time tomorrow. Something
in my content model seems to be adrift.

Some points below. (Btw, am trying to figure out if this is a 2.0 stylesheet
or a universal 1.*,2.* stylesheet as it contains a lot of 1.* references.)

[1] For now, I note that both recordPacking and recordXMLEscaping are
reported in each record. Which I think is good, but I would strongly suggest
to make both optional elements.

[2] Also, I didn't recall exactly why this was included in both
recordDefinition and searchRetrieveResponseDfinition content models:

==
<xs:any minOccurs="0" maxOccurs="unbounded" namespace="##other"
processContents="lax"/>
==

Why? I thought we had the extraRecordData and extraResponseData elements for
extensibility?

[3] Why is recordIdleTime still present with recordTTL added below
extraRecordData. Why isn't -TTL replacing -IdleTime above?

[4] Is version still OK? Am happy with it being optional.

[5] Hmm, and while we're here what about this:

    <xs:simpleType name="versionDefinition">
        <xs:restriction base="xs:string">
            <xs:enumeration value="1.0"/>
            <xs:enumeration value="1.1"/>
            <xs:enumeration value="1.2"/>
        </xs:restriction>
    </xs:simpleType>

So, in summary the one real change needed for now is [1] - making both
recordPacking and recordXMLEscaping optional.

Cheers,

Tony

*********************************************************