Ontology Query Languages for Ontology-Based Databases - Data Warehousing Design and Advanced Engineering Applications - page 229

Defined by Gruber as “an explicit specifica-
tion of a conceptualization”, ontologies have been
proposed to explicit the semantics of data. They
allow to describe, in a consensual way, the relevant
concepts of a given application domain. Thus, the
idea to describe a data warehouse with ontologies
in order to make explicit the semantics of data
stored has emerged (Xuan et al., 2006). This idea
is concretized by introducing an ontological layer
in a data warehouse. This layer can be used by
decision makers to express queries using business
concepts they are used to manipulate. However,
existing query languages such as SQL or OQL
have not been designed to exploit ontologies. For
this purpose, a new generation of query languages,
called ontology query languages , has emerged.
The goal of this chapter is to provide an up to
date survey on the capabilities of existing ontology
query languages to manage databases extended
with ontologies we call ontology-based databases
(OBDBs). Previous surveys on ontology query
languages (Bailey et al. 2005; Haase et al. 2004)
have focused on the capabilities of Semantic Web
query languages to manage RDF-Schema ontolo-
gies and data whatever be the used storage system.
In this chapter, we propose to complete these
previous surveys by taking a database-oriented
point of view. Thus, this chapter discusses two
representatives of Semantic Web query languages,
namely SPARQL (Prud'hommeaux and Seaborne,
2008) and RQL (Karvounarakis et al., 2004) as
well as two representatives of database-oriented
ontology query languages namely, Oracle exten-
sion to SQL (Chong et al., 2005; Wu et al., 2008)
and OntoQL (Jean, 2007; Jean et al., 2006) .
This chapter is organized as follows. In sec-
tion 2, we present our point of view on domain
ontologies and their classification. We conclude
this section by describing our proposed exten-
sion of the traditional ANSI/SPARC database
architecture with ontologies. Requirements for an
exploitation language of this proposed architecture
are then defined in section 3. These requirements
are used to compare some existing Semantic Web

Query Languages and database-oriented ontology
query languages in section 4 and 5. Finally, sec-
tion 6 concludes this comparison and introduces
future work.

DOMAIN ONTOLOGIES AND THEIR
APPLICATION TO DATABASE

In this section, we present the database-oriented
point of view on ontologies we take (Jean et al.,
2007b; Fankam et al., 2008) to compare existing
ontology query languages.

Definition and Classification
of Domain Ontologies

Several definitions have been proposed for an on-
tology (Gruber, 1993; Guarino, 1998). In our work
a domain ontology is “a formal and consensual
dictionary of categories and properties of entities
of a domain and the relationships that hold among
them”. This definition emphasizes three criteria
that distinguish ontologies from other models used
in Computer Science. An ontology is:

1. formal : it is based on a logical axioms and

may be processed by computers; so check-
ing consistency and performing automatic
reasoning are made possible;

2. consensual in a community, i.e. several

members have agreed upon the concepts
represented in the ontology;

3. has the capability to be referenced. A univer-

sally unique identifier can be used to define
the semantic of a piece of data, whatever are
the modeling schema of the ontology and
the data model.

All ontologies are not similar. We distinguish
the three following categories:

• Conceptual Canonical Ontologies

(CCOs) provide concepts definitions