Extensions on the registry

[Guy Rixon]

On Tue, 12 Apr 2005, Ray Plante wrote:

> On Tue, 12 Apr 2005, KevinBenson wrote:
> > Does everybody agree that we should be able to on Full Registries store all
> > Resources including extension Resources?
>
> My current take, personally, is that Full registries *should* store all
> records, even though they contain extensions; however, I'm not sure I
> would go as far as "must", since this has potentially substantial
> implementation issues.  The next question is, what does this do to the
> meaning of "Full"?  If a registry does not include extensions, is it
> allowed to declare itself full, or can we live with a fuzzier definition?
>
> > Would it be okay if certain Full Registries would not be able to query on
> > them and possibly never query on them (them being the extension part of the
> > schema's)?
>
> Yes.

Doesn't it depend on what you mean by query?

If "query" means "extract metadata from resource documents, transform and
present to user as XHTML", then clearly a given registry can't do this
properly for unknown extensions. (Nominating XSLT stylesheets in processing
instructions might help.)

If "query" means "return the resource document matching a given IVOID" then
that works with or without extensions.

If "query" means "return XML fragments extracted from resource documents
according to some query language such as XQuery or ADQL" then it's less clear.
My guess is that an XQuery or XPath query would work even on
resources-with-extensions and that an ADQL/SQL query might have problems.

I think we should mandate the return-entire-document-matching-IVOID query for
all registries even in the presence of extended documents. (I'd also prefer
XQuery over ADQL for the structured query, but that's a separate issue.)

Cheers,
Guy

Guy Rixon 				        gtr at ast.cam.ac.uk
Institute of Astronomy   	                Tel: +44-1223-337542
Madingley Road, Cambridge, UK, CB3 0HA		Fax: +44-1223-337523