[xquery-talk] SPARQL to XQuery transformation

[Rob Gonzalez]

It's Turing-complete, so no, there is no inherent disadvantage to
being purely functional, but it does have the advantage that
computation over large amounts of data can be more easily
parallelized, which is a huge advantage for XQuery in database
systems.

To your other point: inferencing isn't done as part of SPARQL
syntactically; that's part of the underlying data model and query
resolution.  I see no obvious reason, though maybe one exists, why
inferencing can't be done on an XML infoset (which is a directed
graph) in the same way that it's done on RDF data (which is a directed
graph).  I think it would be interesting to try to use OWL to describe
XML (OWL isn't specifically about RDF; it's just about things) to
prove this out (indeed, the fact that RDF can be serialized as XML
makes me think this wouldn't even be that challenging an exercise).

best,
-Rob


On Tue, Apr 28, 2009 at 8:32 PM, David A. Lee <dlee at calldei.com> wrote:
> I don't believe its impossible either, but I was thinking, as a purely
> functional language (if that makes a difference?)
> Is there any set of computations that xquery couldn't theoretically execute,
> that say a procedural language could ?
>
>
> Michael Kay wrote:
>
> It might be *possible* to do in XQuery but its not a simple
> "translation" of one query language to another, it would take
> writing an inference engine in xquery.  quite a project.
>
>
> If you can write an interpreter for a query language then you can write a
> compiler for it. And compiling into XQuery is probably easier than compiling
> into machine code. So as far as I can see, it doesn't look intrinsically
> impossible at all.
>
> Michael Kay
> http://www.saxonica.com/
>
> _______________________________________________
> talk at x-query.com
> http://x-query.com/mailman/listinfo/talk
>
>
> --
> -----------------------------------------------------------
> David A. Lee
> dlee at calldei.com
> http://www.calldei.com
>
> _______________________________________________
> talk at x-query.com
> http://x-query.com/mailman/listinfo/talk
>