Conclusion and Outlook - Enabling Semantic Web Services: The Web Service Modeling Ontology - page 168

chitecture of the Web. It is one of the basic design principles of the Web and
the REST architecture not to provide stateful protocols and resources.
In practice, this means that when SOAP messages are sent and received,
the content of the information is hidden in the body and not addressed as an
explicit Web resource with its own URI. Referring to the content transmitted
via an explicit URI in an HTTP request would allow the content of a message
to be treated like any other Web resource.
A possible approach to address these problems might be to look for alter-
native middleware paradigms to the current Web service technologies. As it
turns out, Tuple Space computing [47] which can be seen as a paradigm for
exchanging data between processes that is very similar to the paradigms of
the Web, shows many of the desired characteristics.
Tuple Spaces have been introduced in parallel-programming languages
such as Linda to implement communication between parallel processes [47].
Instead of sending messages forward and backward, a simple communication
means is provided. Processes can write, delete, and read tuples from a persis-
tent space. Tuple Space computing has very strong advantages; it decouples
three orthogonal dimensions involved in information exchange: reference, time,
and space.
This strong decoupling of all three relevant dimensions has obvious de-
sign advantages for defining reusable, distributed, heterogeneous, and quickly
changing communicating applications. Moreover, a service paradigm based
on Tuple Space computing also revisits the Web paradigm: information is
persistently written to a globally accessible space where other processes can
smoothly access it without starting a cascade of message exchanges.
In fact the Web and Tuple Spaces have many things in common. They are
both scalable information spaces for persistent publication and information
dissemination that (in principle) do not rely on replication. However, tradi-
tional Tuple Spaces provide a flat and simple data model that does not provide
any nesting or linking of data. Hence, tuples with the same number of fields
and the same field order but different semantics cannot be distinguished.
In order to make Tuple Spaces usable for Semantic Web services, it would
thus be an appealing idea to move Tuple Space computing in the direction
of the Semantic Web, by adopting the Semantic Web data model (RDF) and
allowing semantics-aware matching rules, which we could call (borrowing from
RDF triples as the underlying data model) “Triple Spaces”.
Note that these ideas do not make Semantic Web services based on tradi-
tional Web services obsolete. More sophisticated technologies are still needed
in the Semantic Web. Tuple spaces can help to decouple communication; how-
ever, they do not provide answers to data and information heterogeneities up
front. That is still a task for the Semantic Web, as it provides standards to
represent machine-processable semantics of data.
We emphasize also that Triple Space computing is more a complement
to Semantic Web services based on “traditional” Web service technologies
than something aimed at replacement of these technologies. Nevertheless, we