Ontology Query Languages for Ontology-Based Databases - Data Warehousing Design and Advanced Engineering Applications - page 233

canonical - concepts. Thus, the language should
support the exploitation of such concepts.

example, PLIB provide constructors to define
precisely and contextually primitive concepts
while OWL provides conceptual equivalence con-
structors to define a NCCO.All these constructors
are useful to define an ontology according to the
Onion Model. Thus, the language should be able
to integrate constructors coming from different
ontology models. This feature would also help to
support evolution of existing ontology models (for
example, the support of OWL2 (Patel-Schneider
and Motik, 2008)).

Requirement 2: (Definition of non canonical
concepts) The language should support the defi-
nition of non canonical concepts. Queries may
be expressed using these concepts, the query
engine interpreting them in terms of canonical
concepts.

Example. Create the class PostDupont defined as
all messages of the user Dupont.
The last layer of the Onion Model is composed
of the LO part. When an ontology is designed ac-
cording to the Onion Model, its LO part associates
to each concept, one or several terms and textual
definitions. These linguistic definitions allow
human users to understand the ontology and to
reference ontology concepts using their names.
These linguistic definitions are often given in
different natural languages. To make it easy for
members of different countries to use the same
ontologies, the language may support the defini-
tion and exploitation of multilingual LO.

Requirement 4: (Ontology model extensibility)
The ontology model supported by the language
should be extendable to support its evolution
and to manage ontologies defined with another
ontology model.

Example. Add the OWL constructor AllValues-
From (a class composed of all instances whose
values of a given property are all instances of a
given class) to the ontology model supported by
the language.

Requirements Resulting from
Preserving Compatibility with
the ANSI/SPARC Architecture Requirement 3: (Linguistic exploitation) The

language should support the definition and exploi-
tation of linguistic definitions of concepts that may
be defined in different natural languages.

Our proposed architecture extends the ANSI/
SPARC architecture, and thus, it also includes the
usual logical level. Since many applications have
been built using the SQL language to manipulate
data at this level, the language should support
manipulation of data not only at the ontological
level (see previous section) but also at the logi-
cal level. Thus, the language will keep upward
compatibility for existing applications and will
permit to manipulate data at the different levels
of our proposed architecture.

Example. Return the first name and last name of
users with a query written using English names
of concepts and one using French names.
The Onion Model is based on the comple-
mentarity of ontology models. Indeed, many
ontology models have been proposed such as
RDF-Schema (Brickley and Guha, 2004), OWL
(Dean and Schreiber, 2004) or PLIB (Pierra, 2008).
All these ontology models have constructors to
define a CCO composed of classes, properties,
datatypes, and instances. In addition to these
core constructors, each ontology model provides
specific constructors to define an ontology. For

Requirement 5: (SQL compatibility) The lan-
guage should permit the manipulation of data at
the logical level preserving SQL compatibility.