Conceptual Aspects: How Landmarks Can Be Described in Data Models - Landmarks: GIScience for Intelligent Services - page 122

can mean Alexanderplatz in contrast to Unter den Linden , Pergamonmuseum and
Kurfürstendamm . Talking to a nearby friend on the phone to negotiate a meeting
point the reference can mean Alexanderplatz in contrast to Spandauer Strasse , Karl-
Liebknecht-Strasse ,and Karl-Marx-Allee . The contrast principle allows explaining
the localization of objects whilst avoiding geometrically specified positions in
space or characterizing places by boundaries. Places are described as prototypes
or centres, or even dimensionless entities in information space. The spatial extent of
a place is only a refinement, but not core of the meaning of place. This approach fits
well to proximity relations between elements. Proximity to dimensionless entities is
cognitively and computationally more directly accessible than distances between
spatially extended objects, which require more complex concepts such as the
Hausdorff distance [ 25 ] .
Contrast sets always provide a key to choose an appropriate level of spatial
granularity for reference regions. However, in a communication the current contrast
set is only sometimes explicit. In other communication situations the contrast set
has to be concluded from factors such as the location of the person, other locative
expressions, or activities mentioned.

4.2.5 Formal Models for Hierarchies

Two types of hierarchies among landmarks were identified in human spatial
reasoning and communication, spatial granularity and cognitive salience .Inorder
to enable a machine to generate cognitively efficient expressions in human-computer
interaction—expressions that zoom in or out—both kinds of hierarchies need to be
available.

4.2.5.1 Hierarchies and Reference Regions

Configurations of objects, such as the landmarks in our thought experiment in
Sect. 2 , consist of a set of geographic (environmental) objects that typically satisfy
to be non-overlapping, littering the otherwise empty environmental space, and
stationary. The entities representing these objects shall be modelled as conceptual
nodes in information space. Their spatial extent might be represented as a further
property, but is of no relevance for this model. Furthermore this approach facilitates
efficient proximity relations between elements. For modelling proximity we resort
to Delaunay triangulations and their dual, Voronoi diagrams [ 1 , 3 , 10 , 60 , 61 ] .
This way connections between objects are by design limited to nearest neighbors,
which is not necessarily the case in mental spatial representations. While the actual
characteristic of relationships in mental spatial relationships remains opaque, the
general principle of cognitive efficiency would at least support such a design choice
as a good approximation. Theoretical evidence assuming that nearest objects are (at
least stronger) linked in mental spatial representations exists. For example, human