Re: about issue 7

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Well,

the point is that a wsdl element may receive multiple distinct annotations, as this seems largely agreed upon in the
group. I am suggesting to attach related modelReference/schemaMapping pairs so that no inference or scanning is required
to tell which mapping goes with which modelReference.

Btw, after our private discussion last week at the DIP meeting, I am ready to accept that a model reference goes alone,
without a mapping, which still makes sense for discovery for instance.

I am also ready to accept that a model reference has several mappings (one in xslt, one in rdf for instance, maybe one
in Java). I wonder however if the resources that may be used to compute a mapping should be elements of a single
schemaMapping, or define alternative mappings in their own right. Let's call this a detail.

I still do not accept the idea that a single mapping can be used for several model references (even though I can accept
that the same program can be used to map from different schema elements to the same concept).

To summarise, I would favor the fact that we group together a modelReference, and its associated (maybe empty) list of
schemaMapping.

Obviously enough, the "list" can either be explicit (my favorite) or implicit, by letting the modelReferences and
schemaMappings share a common unique iri (or simply recalling in the mappings a "reference" to the modelReference.

Laurent


Jacek Kopecky wrote:
> Laurent, 
> 
> can you please expand a bit on the meaning of "an arbitrary number of
> model references [...] attached to any wsdl element"? I'm sorry, but I
> don't seem to follow the step from pairing modelReference/schemaMapping*
> to arbitrary numbers of model references...
> 
> (*)I changed modelElement to modelReference in the pair, is this
> correct?
> 
> Thanks,
> 
> Jacek
> 
> On Tue, 2006-05-09 at 17:35 +0200, Laurent Henocque wrote:
> As I understand it, the schema mapping is a property of the binding between a swdl element and a modelElement.
> The user needs a mapping to support the claim that the spec is an implementation of the concept.
> 
> If this is correct, I guess that we need the notion of a pair <modelElement/schemaMapping> instead of two separate
> property value tags. That way, an arbitrary number of model references, related to an arbitrary number of ontology
> languages could be attached to any wsdl elemnt.
> 
> I don't believe in a tool that would discover by magic which mappings correspond to which model elements.
> 
> Laurent
> 
> 
> 
>>

- --
*************************************************************************
Laurent Henocque
Maître de Conférences Hdr
tel: +33 6 83 88 20 01
Enseignant à l'Ecole Supérieure d'Ingénieurs de Luminy - Marseille
    http://www.esil.univ-mrs.fr
Chercheur au Laboratoire des Sciences de l'Information et des Systèmes - Marseille
    http://www.lsis.org

clé publique open pgp / open pgp public key :
http://www.esil.univ-mrs.fr/~henocque/0x987E183.pub.asc
************************************************************************
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFEcxYmIF1tz5h+GDARAupbAJ9+2UfU4Q98C0hDwv9FSR6qZEVdEQCfQN9z
eULqBXZnuQ6Tg7eJraDffAw=
=Re78
-----END PGP SIGNATURE-----

Received on Tuesday, 23 May 2006 14:03:42 UTC