Ontology Query Languages for Ontology-Based Databases - Data Warehousing Design and Advanced Engineering Applications - page 245

Table 3. Analysis of the main ontology query languages w.r.t. to the defined requirements

SPARQL RQL Oracle OntoQL

Queries at the ontological level o • o •

Definition of non canonical concepts o o • o

Linguistic exploitation • - • •

Ontology model extensibility o o • o

SQL compatibility - - • •

Access to the conceptual model of data - - - o

Ontology & data definition and manipulation - o o o

Queries on ontologies and on ontologies and data • • • •

Results presented in Table 3 lead us to draw
the following conclusions. The main drawback
of SPARQL and Oracle approaches to query
OBDBs is that they consider all information as
RDF data. As a consequence, when querying at
the ontological level, the semantics of the ontol-
ogy models has to be coded either in the query
interpreter of SPARQL or by a set of deductive
rules in the Oracle approach. Moreover, they
don't provide operators to exploit this semantics
(e.g., an operator to retrieve direct instances of
a class). Another consequence is that the syntax
of these two languages is adapted to query triple
data. As a consequence, an ontological query has
to be decomposed in triples (e.g, (?i rdf:type User)
(?i sioc:name ?n) for retrieving names of users)
which can be tedious for users.
On the contrary, the semantics of RQL and
OntoQL are based on the core constructors of
ontology models. They provide a syntax near
the one of object-oriented query languages. We
think that this syntax is more adapted for ontolo-
gies than a triple syntax because ontologies share
many constructors with the object-oriented data
model.
However, RQL and OntoQL may benefit from
efforts made by Oracle to provide semantics tech-
nologies. Indeed, Oracle provides customizable
and optimized triple storage with capability to
load quickly a huge amount of RDF data. Thus,
Oracle provides the built-in functions to serve as

a scalable storage structure for RQL or OntoQL.
Moreover, Oracle is now equipped with an infer-
ence engine for RDFS/OWL constructs which
will be particularly useful for RQL and OntoQL
to take into account the semantics of OWL. As
a future work we plan to put this observation in
application by implementing OntoQL on top of
Oracle.

REFERENCES

Bailey, J., Bry, F., Furche, T., & Schaffert, S.
(2005). Web and Semantic Web query languages:
A survey. In Reasoning Web, First International
Summer School (LNCS, pp. 35-133).

Brickley, D., & Guha, R. V. (2004). RDF Vocabu-
lary Description Language 1.0: RDF Schema .
World Wide Web Consortium. Retrieved from
http://www.w3.org/TR/rdf-schema

Broeskstra, J., & Kampman, A. (2003). SeRQL:
A second generation RDF query language. In
SWADEurope Workshop on Semantic Web Stor-
age and Retrieval .

Carroll, J. J., Bizer, C., Hayes, P., & Stickler, P.
(2005). Named graphs, provenance and trust. In
Proceedings of the 14th international conference
on World Wide Web (WWW'05) (pp. 613-622).
New York: ACM Press.