Database Reference
In-Depth Information
These people cannot of course be completely neutral or ambivalent to geography
if they wish to use it for data integration. And, it is here that Linked Data and the
Semantic Web may help by providing a common and simple data model based on the
Resource Description Framework (RDF) triple and a machine-interpretable means
to describe the data using ontologies. This common data model provides a founda-
tion that simplifies the process of data integration or linking and is also data neutral:
The same standard applies to bioinformatics just as much as to GI or financial data.
By separating the description of the data from the data itself and exposing it in an
ontology, the Semantic Web approach also allows a third party to understand the
meaning of the data better and to begin an integration or linking process at this
level. The establishment of links between data enables relationships to be explicit
and visible. Hence, the decisions that have been made during the data integration
process are made clearer and are also preserved for others to use.
For the GI community, the adoption of these technologies offers the chance to
open up their expertise and specialisms to a much wider audience. The manner in
which the spatial elements of data are represented and manipulated will be pre-
served, often simply by finding ways to represent existing standards (or the relevant
aspects of them) within a Semantic Web environment. And, once preserved in this
new environment, they naturally become open to many more end users. The general
way in which data is formatted and exchanged now exists within a Linked Data
framework, and the specialist aspects of the data, such as the representation of
geometry, can be enshrined as special datatypes described by the GI community.
Any person who uses Linked Data, irrespective of the person's background, now
has a common representational form for his or her data that will be familiar to
any other person who also uses Linked Data. The GI community can also publish
micro-ontologies that define and describe the vocabularies that are used, in a man-
ner that can be understood by a much wider audience than the GI community alone.
A major barrier to data integration, the problem of varied formats, has therefore
been significantly reduced. The advent of Linked Data now means that GI really
does belong to a much broader community.
GI itself is evolving; there is increasing recognition of what can be pithily sum-
marized as “Place before Space.” By this we mean that a lot of problems can be
resolved without knowing very precise locations or extents as would be tradition-
ally required by a GIS. Place is more concerned with identity, which includes place
names, addresses and postcodes or zip codes, and topologic and mereologic relation-
ships. Location itself is often expressed as a simple point, and the boundary of an
object may not be represented at all.
1
Such emphases are well suited to expression
as Linked Data. We have also seen that some aspects of GI are not well suited to
explicit representation and analysis on the Semantic Web. For example, raster data
can only be referenced by a Uniform Resource Identifier (URI) and cannot be inter-
acted with at all on the Linked Data Web, while analysis and manipulation of vector
geometries is quite limited. This means that GIS will therefore remain an impor-
tant analytical tool, standing alongside the representation of data on the Semantic
Web. Most benefit will therefore be realized by recognizing that the technologies are
largely complementary rather than competitive: GIS are better suited to analysis and
