Database Reference
In-Depth Information
but are required to help describe the core concepts. For example, in a topographic
ontology Woodland will need to make a reference to Trees; without reference to
Trees, it is difficult to describe Woodland, but equally Trees are not central to the
topographic domain. So, while the topographic ontology will need to describe in
detail the concept of Woodland, it will say little or nothing about the concept of a
Tree, at most stating the difference between coniferous and broad-leaved trees. Thus,
in this case
Woodland is a core concept, and a Tree is a secondary concept. Merean
Nature might take a different view as it will also need to be concerned about indi-
vidual tree species, so both Woodland and Tree would be core concepts.
10.3.5 c
onceptUalization
The conceptualization stage is when the hard cognitive work happens and the natural
language glossary is turned into an ontology of formally described classes and prop-
erties in an ontology language such as OWL. This stage will include testing the com-
pleteness and correctness of the ontology by applying the competency questions. One
task that is becoming increasingly important as more ontologies are published is to
check to see whether existing ontologies can be exploited to help build your ontology.
This description gives the impression that these stages are completely sequential;
the reality is very different. Constructing the glossary can significantly overlap with
the lexical analysis as the ontology author may choose to explain terms as they are
discovered. The glossary stage will also be revisited during the conceptualization
stage as new terms are discovered or some terms found to be unnecessary. In partic-
ular, this stage is likely to generate new terms for relationships between the various
classes. So, expect the process to be very iterative.
In the next section, we work through these stages from the perspective of Merea Maps.
10.4
BUILDING THE TOPOGRAPHICÂ ONTOLOGY
OF MEREA MAPS
10.4.1 s
cope
anD
p
Urpose
Defining the scope and purpose can be surprisingly difficult, especially for domain
ontologies. For Merea Maps to define the scope and purpose of its ontology, it first
has to really understand what its purpose is and why it is building the ontology.
This is not an easy question because in this case the ontology will not have a single
specific application. Even where there is an initial application in mind, it may be dif-
ficult to specify scope and purpose simply because different people within the same
organizations may have differing opinions regarding what the scope and purpose
should be. In such cases, the process will help to arrive at consensus and will cer-
tainly highlight differences that might not otherwise be revealed until much later in
the process. Defining the purpose should come first because if you do not know why
you are doing something, the scope is irrelevant. So, why does Merea Maps want an
ontology? Merea Maps sees itself as providing a referencing framework for other
organizations and people, enabling them to use the features that Merea Maps pro-
vides for use within specific applications. For example, Merea Heritage can use the
