Database Reference
In-Depth Information
From a traditional GIS perspective, geometry is the main driver, and the world
is divided into different layers of information, each layer representing a particular
type. Standards are also geometry centric and support the concept not only of
layers but also of features: OO representations of real-world things where informa-
tion about the same object is collected together. Other views also exist; these are
focused less on geometry and coordinates and more on relationships, whether these
are topological, mereological, or lexical. Standards for representing GI have been
developed, but these are largely restricted to use by a subset of the GI community,
are little known outside this community, and may even serve to help enforce its
isolation. GML is an example of this: used by governments, mandated by policy,
but not widely known or applied by those who use GI in its widest sense.
If we were to simplify things and if we just separated the uses of GI into two
groups, however imperfect this may be, then GI can be broadly split into
1. GI for analysis. Here, the emphasis is to generate new information by per-
forming special spatial operations between different datasets.
2. GI for integration. Here, the purpose is to combine different data, either as
an end in itself or as a prerequisite to performing spatial analysis that typi-
cally involves multiple data sources.
Our focus in this topic is to bring together Linked Data and GI. It is therefore
natural that we concentrate on the latter use of GI for integration, as Linked Data has
much to offer in this area, and it addresses the widest audience. In the next chapter,
we start to examine at a high level how the ideas behind the Semantic Web and
Linked Data require us to think differently about GI (and indeed about information
and data generally). This next chapter is thus extremely important as through the
introduction of this different way of thinking about data we provide a foundation for
the rest of the topic. Following from Chapter 4, the topic introduces the more techni-
cal details behind Linked Data and Semantic Web ontologies and shows how these
can be applied to GI.
NOTES
1. Mereology describes whole-part relationships.
2.
Real-world objects
is a GI term used to refer not only to physical geographic features
such as buildings, roads, woods, and rivers but also less-tangible things such as an
administrative area such as a state or county or health district and even something that
may only exist for specific times in the same place, such as a Monday farmers' market.
3. While it is true that European directions more frequently include landmarks for refer-
ence than, say, a set of directions given by an American, it is still notable that they are
wholly absent from the machine-generated directions.
4. In OGC terms, a feature is a digital representation of a real-world object, reflecting a
single perspective of that object and containing some explicit locational information.
5. Strictly, it could be argued that KML also originated with the GI community. KML
was created by Keyhole Incorporated to enable geospatial visualization but was picked
up by many casual Web developers needing to represent mapping in their applications.
This success led Keyhole to be absorbed by Google in 2004.
