Semantic Web Service Usage Tasks in WSMO - Enabling Semantic Web Services: The Web Service Modeling Ontology - page 131

XML-Level Transformations
In this approach, the grounding needs to link the inputs and outputs of the
service to the appropriate XSLT transformations. With these links in place,
the actual run-time process is very simple: assuming that a client, for example,
wants to send a message to a Web service, it serializes its ontological data into
the XML syntax of WSML and then runs the appropriate XSLT stylesheet,
sending the result to the service. A message from the service goes through
the inverse process. First it is transformed, using a reverse XSLT stylesheet,
into XML. The client then parses the result directly into ontological data and
processes it accordingly.
This approach is very simple and uses proven existing technologies, but it
has a notable disadvantage: the XML representation of ontological data con-
tains an unpredictable mixture of hierarchy and interlinking, as ontological
data is not structured according to XML conventions (we say that ontological
data is not native XML data). Therefore, creating robust XSLT transforma-
tions for complex data structures is a considerable task. With simple data,
however, this problem is negligible, and since XSLT processors are readily
available and many XML-savvy engineers have some XSLT experience, this
approach is an ideal initial candidate for data grounding.
Exchanging Ontological Data
The XSLT transformation approach described above makes use of the fact
that Semantic Web service ontology languages can be serialized in XML. In
certain situations, this XML format could in fact be suitable for the WSDL
interface of a Web service when the data will seldom be viewed as actual
XML. For example, this is ideal for quick Semantic Web service prototyping
of Semantic Web services, or when only semantic clients are expected to use
a particular service.
In these scenarios, the WSDL description of the messages will only indicate
that they contain WSML/XML, and the semantic description will say (as it
does regardless of the grounding approach taken) exactly what data goes in
the inputs and outputs. The grounding needs only to link the semantic inputs
and outputs with particular WSDL messages, which is done implicitly by all
the behavior-grounding approaches that we later describe in Section 9.5.3.
This approach does not require any human designer to create grounding
transformations, which indeed saves a significant amount of effort. On the
other hand, XML serializations of ontological data are not native XML data,
and so they may be hard to comprehend or hard to process by XML tools.
Consequently, services that use this grounding approach may not integrate
well with non Semantic Web services.