Database Reference
In-Depth Information
6.7 SUMMARY
This chapter has shown how GI may be represented as Linked Data. It has discussed
identity in terms of both digital identity (URIs) and place names, as well as covering
basic classification and shown how topological and mereological relationships can
be handled. It has also highlighted how there are always choices to be made, and
that the modeler is always faced with balancing conflicting interests—most often
between model richness and complexity. Understanding end-user requirements and
minimizing complexity are the most important factors to take into account when
representing GI as Linked Data.
NOTES
1. This is your preference, no one else's.
2. For brevity, from now on we will omit the statement that maps a succinct name to a URI,
in this case: Ash Down farm (name) has URI mm:/181113.
3. The model allows multiple geometries to be associated with the OGC Feature; here, we
only deal with the default. The representation of other geometries is the same.
4. For completeness, OGC also implemented the Egenhofer nine-way intersection model.
5. In the United Kingdom and Merea, one use of the term
terrace
is to describe a row of
houses or cottages that are physically constructed as a single building, and hence each
house or cottage shares at least one wall with another.
6. There is one caveat that needs to be made clear here: The range and domain of the
OGC properties are geo:SpatialObject. If you do not want your features to be inferred to
belong to this class, then you will have to implement your own properties.
7. If Merea Maps could be sure that semantic Web applications using their data would only
present the rdfs:label to the end user, an alternative here would be to encode the term
intended for human consumption, “next to,” in the rdfs:label instead.
8. From here, we only use the Rabbit notation to avoid needless repetition unless the use of
Turtle is helpful to demonstrate some aspect of the syntax not previously introduced.
9. There are occasions when the topological and mereological properties are strongly con-
nected; this is explicitly recognized as
mereotopology
, for which the properties imply
both part-whole and topological properties.
10. This statement is only required if you are interested in both the connectivity between
roads and the connectivity between places.
