Print

Print
Yes I see.  Well there are two different things we're talking about:
 
(1) the record metadata schema could be  used as the actual record schema to retrieve the record.
 
(2) you want to include record metadata elements from that namespace, but you want to use a different schema to retrieve the actual record.
 
It's only in case (1) where schema validation (RMS schema) is an issue. And you are interested in case (2).
 
In the example ......
 
---------------------------
<record>
   < recordSchema>info:srw/schema/1/dc-v1.1</recordSchema>
       <recordPacking>xml</recordPacking>
    <recordData>
        < srw_dc:dc>
            <dc:title>Thisis a Sample Record</dc:title> </srw_dc:dc>
    < /recordData>
    <recordPosition>1</recordPosition>
    <extraRecordData>
     <rel:rank xmlns:rel="xmlns:rob="info:srw/extension/2/relevancy">0.965</rel:rank
   < /extraRecordData>
< /record>
________________________________
 
... the record schema is dublin core. That's independent of the extraData. 
 
The content of the extraData element is an extension, whose URI is info:srw/extension/2/relevancy. (In this case the URI identifying the extension is doubling as the namespace URI for the extension. However, though perfectly legal, this isn't necessarily a good practice, and not all extensions do this. )  The "2" means that it was defined by Rob, and he called it "relevancy". Unfortunately it's not registered, it's just shown as an example, but if it were registered then the definition would say "put the normalized rank of this record in the element <rank>" ("normalized' meaning a value between 0 and 1).
 
Note that an extension doesn't have to define a formal schema, it defines a namespace, some elements, and rules for how the xml is to be constructed (so an "informal" schema).
 
So if you want to include a set of metadata along with a record in this fashion, you  would define an extension is a similar manner.
 
Is this making more sense now?
 
--Ray
 
----- Original Message -----
From: [log in to unmask] href="mailto:[log in to unmask]">Martin Morrey
To: [log in to unmask] href="mailto:[log in to unmask]">[log in to unmask]
Sent: Tuesday, March 20, 2007 1:12 PM
Subject: Re: Record Metadata Schema (was Re: "collection" context set)

Sorry, I realise I gave the wrong URL for the Record Metadata Schema,
what I meant was http://srw.cheshire3.org/schemas/rec/1.0/.

I am trying to reconcile the example given at the bottom of this page,
with the examples in http://www.loc.gov/standards/sru/extra-data.html
and  http://www.loc.gov/standards/sru/sru-spec.html#resultmodel

At a guess, an example using some of the extra "history" fields in the
RMS might look like this:

<record>
   <recordSchema>info:srw/schema/1/dc-v1.1</recordSchema>
   <recordPacking>xml</recordPacking>
   <recordData>
      <srw_dc:dc>
         <dc:title>Thisis a Sample Record</dc:title>
      </srw_dc:dc>
   </recordData>
   <recordPosition>1</recordPosition>
   <extraRecordData>
      <rec:history xmlns:rec="http://srw.o-r-g.org/schemas/rec/1.0/">
         <rec:created>2002-12-09T12:00:00</rec:created>
         <rec:lastModified>2002-12-09T12:30:00</rec:lastModified>
      </rec:history>
   </extraRecordData>
</record>

The use of "rec" as the namespace prefix is arbitary, the value of the
namespace, "http://srw.o-r-g.org/schemas/rec/1.0/", is specified by the
RMS document.

However, I don't think the example above would validate against the
actual XML Schema for RMS,
http://srw.cheshire3.org/schemas/rec/1.0/rec.xsd, which requires the RMS
fields are contained directly within a root "record" element, not
"extraRecordData".

Hope that makes some kind of sense!

Martin


Ray Denenberg, Library of Congress wrote:
> Martin -- If what you're trying to do is include record level metadata
> with a response record it's a fairly straightforward process.   Use the
> extraRecordData parameter which you'll see described in the parameter
> table in the result set model section (that you cited), more fully
> described at http://www.loc.gov/standards/sru/extra-data.html.
> 
> --Ray
>
>     ----- Original Message -----
>     *From:* Martin Morrey <mailto:[log in to unmask]>
>     *To:* [log in to unmask] <mailto:[log in to unmask]>
>     *Sent:* Tuesday, March 20, 2007 8:54 AM
>     *Subject:* Record Metadata Schema (was Re: "collection" context set)
>
>     I think we could also make use of elements from the Record Metadata
>     Schema, http://www.loc.gov/standards/sru/sru-spec.html#resultmodel
>
>     Do you have any advice on how to combine elements from the RMS with the
>     standard SRU Result model,
>     http://www.loc.gov/standards/sru/sru-spec.html#resultmodel?
>
>     At first glance, its not quite clear to me what the relationship
>     between
>     these two things is.
>
>     Many thanks,
>
>     Martin
>
>     Rob Sanderson wrote:
>      > On Mon, 2007-03-19 at 12:28 +0000, Martin Morrey wrote:
>      >> Rob Sanderson wrote:
>      >>>> The aim of Intrallect's "collection" context set is to allow
>     queries to
>      >>>> be limited to sub-collections within a repository.
>      >
>      >>> How would that differ from
>      >>> collectionName and collectionIdentifier in the record metadata
>     context
>      >>> set?
>      >
>      >> I think it is the same thing.  Glad I asked!
>      >
>      > I think so too :)
>      >
>      > Rob
>
>
>     --
>     ----------------------------------------------------------------------
>     Martin Morrey, Product Director, Intrallect, http://www.intrallect.com
>     [log in to unmask] <mailto:[log in to unmask]>, Tel: +44
>     870 234 3933, Fax: +44 1506 505 117
>     ----------------------------------------------------------------------


--
----------------------------------------------------------------------
Martin Morrey, Product Director, Intrallect, http://www.intrallect.com
[log in to unmask], Tel: +44 870 234 3933, Fax: +44 1506 505 117
----------------------------------------------------------------------