[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: XML Data Modellling/Linking (was RE: AfterXQuery

• To: "Michael Kay" <michael.h.kay@n...>,<davep@d...>, <xml-dev@l...>
• Subject: RE: XML Data Modellling/Linking (was RE: AfterXQuery, are we done?)
• From: "Hunsberger, Peter" <Peter.Hunsberger@S...>
• Date: Mon, 25 Oct 2004 09:48:50 -0500
• Thread-index: AcS4aLXVsgzsIodSTSOWkgd+UvfwKQACwEEgAIrzcIA=
• Thread-topic: XML Data Modellling/Linking (was RE: AfterXQuery, are we done?)

Michael Kay <michael.h.kay@n...> writes:

> > Your first message in this thread intimated you had ideas 
> on how XML 
> > markup might  be added to 'style' (or otherwise process) 
> such items to 
> > move towards the ideas of XLINK....
> 
> You and others seem to have misread my intent. I have no 
> answers to bring to the table, only a strong feeling that 
> there is a piece of the jigsaw that is still missing. XML is 
> a hierarchic data model but the world is a network. There are 
> lots of possible ways to fill in the missing links [sic], but 
> none of them feels very satisfactory (for example, many of 
> them only work for intra-document relationships).
> 
> I'm not even comfortable that the hierarchic relationships 
> should be special. Why can't we have multiple hierarchic 
> views of the same network? Why do all my queries have to 
> change depending on whether my footnotes are inline, 
> out-of-line referenced by IDREFs, or in external documents 
> referenced by URI? What happened to the old doctrine of data 
> independence?
> 

Fair questions, but I think the reality is that you are going to be
forced to pay the price one way or the other: the question of a single
easily traversable representation or a hard to traverse graph/network
translates into the question of high processing costs vs. high storage
costs (once more).  

Even today, if you don't mind paying for the processing costs you can
get structural independence, just uniquely name every element and every
attribute and then use something like: //uniquename.  Somehow I don't
think that's what you want. The alternative of providing translations
for multiple views also works, but you still have to know what to look
up to match the particular query against the particular view, now you're
diving back into ontologies and the like. 

If I truly had to solve the issue of a universally (hah, such ambition!)
addressable network model for some broad domain then given the current
state of the art (and my limited understanding of the alternatives) I'd
have to choose some underlying RDF like storage mechanisim and some OWL
like mapping layer.  That gets me a back end and a middle layer, but now
what? As I said before it's either a WS something or other query wrapper
or a REST/xQueryX approach. The former has some standards arriving on
the table, but I don't think the latter has anything approaching
maturity (read universal applicability) yet?

• Prev by Date: Re: XML Data Modellling/Linking (was RE: Afte rXQuery, are we done?)
• Next by Date: RE: After XQuery, are we done?
• Previous by thread: Re: XML Data Modellling/Linking (was RE: AfterXQuery, are we done?)
• Next by thread: Conformance testing and the XML-test-suite
• Index(es):
  • Date
  • Thread