On Wed, 2007-03-21 at 15:27 +0000, Martin Morrey wrote:
> I was inspired to ask about this by Rob Sanderson's reference to the
> "record metadata context set",
> http://srw.cheshire3.org/contextSets/rec/1.1/, in a response to a
> previous post. I kind of thought Rob would have something to offer on
> what appears to be the equivalent "Record Metadata Schema" (RMS),
> http://srw.cheshire3.org/schemas/rec/1.0/. I am specifically interested
> in using some the elements defined in the RMS schema in a standard
> SRU/SRW response.
What I need to do is create the real equivalent :) That schema is the
equivalent of the rec 1.0 context set from before we (as a
community/list) hammered on it a lot to get it up to speed with the
sorts of things that were really needed.
I'm also more than happy to let someone else produce the actual xsd
file.
> The main thing I don't understand is what the relationship is between
> the RMS schema, as defined at http://srw.cheshire3.org/schemas/rec/1.0/,
> and the "record" part of the standard SRU/SRW result model,
> http://www.loc.gov/standards/sru/sru-spec.html#resultmodel, if any.
> Both have a "record" element, but the similarity seems to end there.
Yes, that's about it.
The idea is/was that if you wanted to see the record metadata for an
object, you could do a search with recordSchema set to the metadata
schema's identifier -- then the contents of recordData would be an
instance of the metadata schema. However the context set and schema are
out of sync.
Equally, it would be possible to define an extension to include the
schema in the extraRecordData section, as you've described already.
> So, are the data elements in the RMS schema meant to be used:
> (a) collectively as an alternative to the "record" part of the SRU/SRW
> result model
> (b) individually as extensions to be included in the "extraRecordData"
> field of the standard response
> (c) either of the above
> (d) none of the above, i.e. something else entirely
(c)
> If (a):
> - what parameter should be used in an SRU/SRW request to require the use
> of the RMS schema in the response?
recordSchema=info:srw/schema/2/rec-1.0
or
recordSchema=rec
if that's the assigned short name in the server's explain file.
> - how should specifically the RMS schema be fitted into the SRU/SRW
> result model
As the contents of recordData:
<record>
<recordPacking>xml</recordPacking>
<recordSchema>info:srw/schema/2/rec-1.0</recordSchema>
<recordData>
<rec:record xmlns:rec="http://srw.o-r-g.org/schemas/rec/1.0">
<rec:size>102003</rec:size>
...
</rec:record>
</recordData>
</record>
(should fix that namespace too)
> If (b):
> - why does the RMS schema include the "record" element, that already
> exists in the standard response format?
> - what form would the response take with these specific extensions (e.g.
> the sample I included in my previous post)?
Yes. However if the extension said: Please give me any metadata you
have, then you would want to have some structure to that response rather
than including it directly in the extraRecordData. So:
<extraRecordData>
<rec:record xmlns:rec="http://srw.o-r-g.org/schemas/rec/1.0">
<rec:size>102003</rec:size>
...
</rec:record>
</extraRecordData>
:)
Rob
|