Is it too late in the day to think about introducing a metadataSchema
element to ZeeRex's schemaInfo, and a metadataSchema parameter
(analogous to recordSchema) in SRU, and dropping the "x-info-99-
metadata..." thing entirely?

This would seem the cleanest way to distinguish between the two in
explain, and the most consistent within SRU.

Of course it probably raises further questions like: "do we allow record
and metadata to be requested simultaneously?", "can we use the existing
diagnostic for unknown schema, or do we need a new one?"...

John



On Tue, 2007-05-08 at 10:06 -0500, Ray Denenberg wrote:
> More that that though...
> 
> If you have a format that can use to present record data, or you can use it
> to present record metadata, then if you specify that schema in the request,
> how is the server to know which of the two you're asking for?  Hence my
> suggestion to assign two different identifiers for the two different roles.
> Does that not make sense?
> 
> And that's one of two reason. The other reason is to be able to distinguish
> in explain which of the two you support, if you support one but not the
> other. 
> 
> --Ray
> 
> -----Original Message-----
> From: SRU (Search and Retrieve Via URL) Implementors [mailto:[log in to unmask]] On
> Behalf Of Dr R. Sanderson
> Sent: Tuesday, May 08, 2007 10:19 AM
> To: [log in to unmask]
> Subject: Re: Record Metadata Schema
> 
> I'll claim that.
> 
> I certainly wouldn't want to try to display a record which was natively in
> (say) SVG in a record metadata schema... I'd want to display the metadata
> about that record in that schema.
> 
> Rob
> 
> On Tue, 8 May 2007, LeVan,Ralph wrote:
> 
> >> -----Original Message-----
> >> From: SRU (Search and Retrieve Via URL) Implementors
> > [mailto:[log in to unmask]]
> >> On Behalf Of Ray Denenberg
> >>
> >> Two. If you list marcXML in explain as a supported record schema, but
> > you
> >> only support one of these, similarly you need a way to distinguish,
> > and
> >> similarly, assigning distinct identifiers is a clear and painess way.
> >
> > But you clearly support both schemas for your records, so they should 
> > both be listed in the explain record.
> >
> > Are you claiming that there are schemas that you would use for 
> > metadata that you couldn't use for the data itself?
> >
> > Ralph
> >
-- 
John Harrison <[log in to unmask]>
University of Liverpool Library