RE: requesting XML records

On Thu, 27 Mar 2003, Matthew Dovey wrote:

> > >> Yes.  I agree this is an issue.  What I think you want to do is 
> > >> request a metadata record that conforms to any 
> > 'application profile' 
> > >> that is based on 'simple DC'.
> > 
> > I think this is a good articulation of Theo's problem, and 
> > I'd like Theo either to confirm this, or explain why I am 
> > wrong.  If we can all agree that this is what you really want 
> > to say, Theo, then we have a better chance of actually 
> > solving your problem.
> > 
> > (Background for those who are not on the ZNG list: we've had 
> > much the same discussion with Theo on that list, and the 
> > rather depressing conclusion seemed to be that Theo wanted to 
> > ask for "records that are either in schema A or 'reasonably 
> > close'", which of course is not something you can ask a 
> > computer to judge.  If it turns out that "based on simple DC" 
> > is an adequate paraphrase for Theo's needs, we may have some 
> > mileage after all.)
> 
> The solution I was proposing on the ZNG list - which I still think would
> work, would be to specify four record syntaxes that we can ask for:
> 
> Strict simple DC - i.e. must use all the DC elements and only those.
> Strict qualified DC
> Loose simple DC - i.e. can include other non-DC elements
> Loose qualified DC
> 
> (I think the names I used were different).
> 
> Hopefully the semantics of this could be such that a server could
> validly return a strict simple DC record when requested for all these
> formats, i.e. use of the others is a hint to the server that the client
> is prepared to except a richer record.

And presumably each of these would have an appropriately strict or loose
XML schema associated with it - so, applications would still phrase their
requests by providing the URL of the XML schema associated with each of
the 4 'labels' above?

This sounds reasonable to me... but there is one thing that might be worth
thinking about...

Is an application profile that uses 1 DC element and 50 IEEE LOM elements
a 'loose simple DC' application or a 'loose IEEE LOM' application or both
or either? ;-) The point being that the naming above (and Theo's DCX name)
takes a very DC-centric view of the world.  Does this matter?

Andy
--
Distributed Systems, UKOLN, University of Bath, Bath, BA2 7AY, UK
http://www.ukoln.ac.uk/ukoln/staff/a.powell       +44 1225 383933
Resource Discovery Network http://www.rdn.ac.uk/

Received on Thursday, 27 March 2003 07:46:57 UTC