FW: [Moderator Action] RE: XQuery as a multidatabase language

Forwarding this reply of Peter (bounced because he used a 
different email address, now added to the accept list).

-M

-----Original Message-----
From: Peter Fankhauser [mailto:fankhaus@ipsi.fraunhofer.de]
Sent: Friday, September 12, 2003 12:54 PM
To: 'andreas jacobsen'; www-ql@w3.org
Cc: Lehti@ipsi.fhg.de
Subject: [Moderator Action] RE: XQuery as a multidatabase language




My colleague Patrick Lehti (lehti@ipsi.fhg.de) is working on this.

In principle XQuery (and/or XSLT) can of course resolve the entire range
of differences
(naming, scale, structure) between semantically equivalent but different
representations of information, simply because XQuery is a
computationally complete functional
programming language. By means of user defined functions as the poor
man's version of
the database notion of views one can even encapsulate a good deal of the
underlying
complexity.

However, using XQuery for the entire job can result in fairly complex
(and difficult to optimize) integration architectures. Also, evolution
(addition of new datasources) can get a bit tedious. Patrick is
therefore
exploring an approach (a variant of the "local-as-view-approach") that
uses
OWL as goal schema. Local schemas are declaratively mapped to this goal
schemas by means of OWL-assertions (same-class as etc.),
representational
differences between equivalent schema constituents are overcome by means
of
an XQuery-like language for OWL instances. Thereby semantic equivalence
is
sort of decoupled from the underlying implementation that resolves
representational differences. This approach should allow for more
modular and evolvable integration architectures, and bears lots of
interesting possibilities for consistency checking, optimization, etc.

Hope this helps,

Peter      

> -----Original Message-----
> From: www-ql-request@w3.org [mailto:www-ql-request@w3.org] On 
> Behalf Of andreas jacobsen
> Sent: Freitag, 12. September 2003 12:17
> To: www-ql@w3.org
> Subject: XQuery as a multidatabase language
> 
> 
> 
> Hi all,
> 
> I'm currently working on my master's thesis at the University 
> Of Bergen, 
> Norway. In my thesis I (try to) explore the capabilities of 
> XQuery in a 
> multidatabase context. This does not merely mean accessing data from 
> several XML documents, but also resolving the semantic 
> heterogeneities 
> possibly (and most likely) occurring when attempting this.
> 
> I am interested in contact with anyone who's working on 
> similar problems or 
> others who think they could have useful input in this area.
> 
> My current conclusion is that XQuery can access several 
> XML-documents, but 
> does that it does not offers any means to resolve differences such as 
> differences such as homonyms and synonyms in the mark-up.
> 
> Regards
> Andreas Jacobsen
> University Of Bergen, Norway
>

Received on Friday, 12 September 2003 09:40:18 UTC