Re: SPARQL in loweringSchemaMapping

Dear Pierre,

while the usage guide shows the use of SPARQL followed by XSLT, this
approach is only recommendable for simple things. I now generally
recommend just XSLT, and I see that's what you're doing.

But just doing XSLT only works if you know what the data serialization
will look like, or if you use some XSLT library (like [1]) to access the
RDF data. You see, I don't expect you'll use an "http-accessible
ontology which has already been serialized in a specific RDF format" as
the data to be sent to the Web service. Lowering transformations work on
data, not on the ontology.

Knowing how the data is formed, one can make assumptions on what the
serialization will look like. Such approach may prove fragile in light
of future changes to the serializer, and it may break in non-intuitive
or hard-to-spot ways because the assumptions are not written down or
verified. That's why the SAWSDL UG uses SPARQL to preprocess the RDF,
and that's why in [1] I suggest using a preprocessing XSLT step and a
set of helper functions.

Basically, XSLT can work for you, and [1] gives you tools for XSLT 2
that should be helpful to make XSLT work for lowering.

Hope it helps,
Jacek

[1] http://www.wsmo.org/TR/d24/d24.2/v0.1/20070412/rdfxslt.html



On Thu, 2007-06-28 at 17:57 +0200, Pierre.CHATEL@fr.thalesgroup.com
wrote: 
> Hi List,
> 
> we are trying to implement SAWSDL in our own System of System but are facing
> problems with data adaptation after semantic web service selection. We are
> actually using the method you indicated in the SAWSDL usage guide (26
> January 2007 Draft) but without SPARQL (only xslt files and the domain
> ontology as a common data model between service providers and consumers
> datatypes).
> 
> Could you give us more information on why and when SPARQL is needed to
> perform service invocation when a common OWL ontology is already available
> for providers and consumers to see. We understand that there could be "many
> variations of RDF semantic data representations" but that's probably not the
> case if there is an http accessible ontology which has already been
> serialized (in a specific RDF format).
> 
> Thanks in advance !
> 
> --
> Pierre Chatel
> THALES Land and Joint Systems
> DLJ/FR/OPS-F/SAT/SC2
> 1-5, Avenue Carnot
> 91883 Massy CEDEX - France
> Tel +33 (0)1 69 75 30 57
> 
> 
>

Received on Thursday, 28 June 2007 16:11:25 UTC