Database Reference
In-Depth Information
clear whether Ash Fleet Farm Estate is part of the village of Ash Fleet or not. It is
within the parish, but the parish is an administrative division, and we should not mix
geographies. Put simply, in the vernacular, the farm is ambiguously defined as either
part of the village or not depending on differing viewpoints. Further, the reason for
the ambiguity is often because it simply does not matter that much; if it did, people
would have resolved the issue. Another problem is that if we do take a geographical
approach, then do we include hills, valleys, and other natural topographic features?
And if we do, how do we relate the mereological to the man-made features? These
questions can only be properly answered if the context is known, but again there is a
danger of mixing different geographies. Therefore, we can conclude that the overuse
of merelogical properties not only is unwieldy but also exposes all sorts of issues that
are difficult to model consistently and offer little advantage to the modeler. So, within
geography what we frequently find is that mereological properties are most useful
for defining the makeup of well-defined places such as farms, hospitals, and so on.
In doing so, it is also important that we should ensure that the description reflects a
single discrete geography, and that we do not confuse topological and mereological
properties
9
or indeed other properties such as ownership.
So, how should we represent mereological properties? Unlike topological prop-
erties, by and large mereological properties are simpler. In fact, the two used most
will be “part of” and “comprises” or “has part.” We define these much as we defined
“next to”; again, we cannot express this using RDFS alone. Most specifically, what
is often useful is to be able to infer that if A is part of B and B is part of C, then A is
also part of C. How this can be achieved is described in Chapter 10, which describes
the application of the OWL ontology language to the domain of geography.
6.6.3 N
eTwork
T
opology
Network topology is the expression of properties that exist between elements of a
network of some form or other. In geography, the most common networks are those
of roads, rail, and rivers. As Linked Data is structurally a graph (or network), it is
able to naturally represent network topology.
In terms of the topological properties themselves, they are mostly concerned with
connectivity (Is A directly connected to B?) and flow (If A is connected to B, can
I travel from A to B in both directions or not?). Other properties can affect the ease
of flow (flow rate, travel times, etc.; the nature of the connection), whether it is a
metaled road, a bifurcating stream, and so on and perhaps things like the number of
connections in the case of multilane roads or multitrack railways. We initially use
the road network between Medina and the village of Isis as an example (
Figure 6.5
).
Here, there are two roads that link the two settlements: the Old Medina Road that
goes via the hamlet of Norton Soke; and the newer Medina Road that goes directly
to Isis.
The simplest way to express the fact that you can travel between the two
settlements is to say something like the following:
Medina is connected to Isis.
