Ontology Query Languages for Ontology-Based Databases - Data Warehousing Design and Advanced Engineering Applications - page 232

Figure 3. Running example

REQUIREMENTS FOR
AN ExPLOITATION
LANGUAGE OF OBDB

tion are possible. Our architecture adds one more
independency, representing both the conceptual
and the ontological description of the represented
data. Thus, exploitation of data independently of
their logical schema is possible. To illustrate our proposed requirements for an

exploitation language of the OBDB architecture
presented in the previous section, we use the
ontology presented as a graph on Figure 3. This
example is an extract of the SIOC ontology ( http://
sioc-project.org/ ). This ontology describes the
domain of on-line communities.A forum (Forum)
is hosted by a site (Site) administered by a modera-
tor (has_moderator). Users (User) can subscribe
to forums (subscriber_of) and create messages
(Post) on these forums (has_container). Several
answers may be provided for a given message
(has_reply).

Requirement 1: (Queries at the ontological
level) The language should allow to express que-
ries at the ontological level independently of the
logical representation (schema) of data.

Example. Retrieve all (direct and/or indirect)
instances of the class Item.
Another fundamental characteristic of the
ANSI/SPARC architecture is to define an external
level. This level defines external schemas (views)
that reflect users perception of the application
domain (e.g., woman in place of person whose
gender is female). These views can be used to
define queries. The query engine rewrites them
according to the logical model of data. In our ar-
chitecture, the NCCO layer of an ontology provide
views at the onlogical level. It allows each user
to represent its own perception of the applica-
tion domain by defining non canonical concepts
expressed in terms of other - canonical or non

Requirements Resulting
from the Onion Model

One of the most important characteristics of the
ANSI/SPARC architecture is to separate the physi-
cal and logical representation of data. Defining,
manipulating and querying data at the logical
level independently of their physical representa-