>1. Am I right in assuming that SRU can (or will) only assume one schema per
>record?
>What do I do with SRU? Do I have to choose one schema over another, or can
>one employ multiple schemas as shown in sample document here?
Going by your example, let me rephrase your question:
Can you have more than one element directly within <sru:recordData>?
To which the answer is no -- it should be a single XML document.
If you want to have two schemas returned in one go, I'd recomment using
a very simple wrapper schema that imports the other schemas.
eg:
<recordData>
<my:wrapper>
<dc:dc>
...
</dc:dc>
<prism:prism>
...
</prism:prism>
</my:wrapper>
</recordData>
If you want a more heavy weight approach there are schemas such as METS
which add more structure to this sort of thing, rather than just an
ad-hoc wrapper schema.
>2. Going with DC, is it OK to split authors into multiple dc:creator fields?
Sure :) Preferable, in fact.
>3. The default SRU query returns an SRU searchRetrievResponse document (as
>below). However, we still want to provide end users with native X/HTML
>result sets and would also prefer to publish a standard search query syntax
>(e.g. SRU), even if we recognize that this is somewhat outside the spirit of
>SRU by pursuing the fiction of an intermediate server which "unpacks" a
>fictional searchRetrievResponse and "repackages" the XML result records into
>an X/HTML page. Assuming that this is a viable option, I presume the way to
>incorporate a switch is to use the SRU "extraRequestData" parameter on the
>query, but I am a bit uncertain what value to assign it.
Yes, but it's not a parameter itself.
The way the extra request data in SRU works is that every parameter
beginning with x- is treated as an extension (in the same way as X-
headers in HTTP etc)
so you could have:
http://www.nature.com/sru?version=1.1&...&x-nature-html=1
It wouldn't be SRU, as previously discussed, but it would be OpenSearch
and only one very very small step away from SRU.
>Thanks again for any help. We are trying to arrive at some conformancy in
>the search space, so please be gentle to the village idiot.
That you're coming to the table rather than doing your own thing is
99.999% of the hard work :) We're happy to help where possible.
Rob
|