Database Reference
In-Depth Information
to raster and vector data, while the Semantic Web is better at the representation of
data in a form that maximizes reuse, interaction, and integration.
The following are points to remember:
• GI is really a broad church and exists in many different forms.
• GI is an important element in the integration of data;=.
• GI standards can be encapsulated within the Semantic Web with vocabu-
laries described by micro-ontologies and standardized geometric repre-
sentations expressed as special datatypes.
•
GI itself is evolving, and great recognition is being given to “Place.” Place
is well suited to representation as Linked Data.
•
Linked Data and the Semantic Web can help GI become a more integrated
part of the wider information community.
•
Linked Data and the Semantic Web do not replace the need for GIS. They
cannot match the analytical properties of a GIS, and they are not well suited
to all forms of geographic data, such as raster. Many aspects of the older
GIS technologies are therefore complimentary to the Linked Data approach.
11.3 AN OPEN WORLD
For those not previously aware of the open world assumption, the implications of its
application are one of the most important messages in this topic. The open world
assumption pre-dates the Web, but its way of looking at the world is ideally suited
to dealing with data held on the Web. The crisp boundaries that define the scope of
a conventional database are simply not present on the Web, and we therefore cannot
assume that if we cannot find a fact it cannot be true and so must be false. The Web
itself is also littered with contradictions: “facts” posted by different publishers that
differ from each other. This world we experience on the Web is also found in every-
day life. Conventional databases essentially ignore these difficulties to provide a
consistent and managed view of a problem domain. This strategy is entirely sensible,
especially where there is an internal focus within an organization or across a closed
group of like-minded organizations. The strategy is less successful when applied to
a broader population, where a diversity of perspectives exists, that cannot, for legiti-
mate reasons, be coerced into a single shared worldview. Indeed, a characteristic of
the Web, and life in general, is that there is no one single worldview. On the Web,
strategies based on the open world Assumption fare better as the goal is no longer to
share a single view but to take part in a discourse among many different viewpoints.
One way in which this difference manifests itself is the way in which we deal with
information that we had not considered at the outset. In the closed world approach,
we are obliged to determine in advance all the types of data that are of interest.
If we are dealing with a road, then perhaps we will identify that we need to hold its
name, the identities of the roads to which it connects, and its position and geometry.
We may also create a constraint: that all roads must have a name. Once done with
modeling, we can go about populating the database with values representing the
roads in which we are interested. And, we can of course do the same when apply-
ing the open world assumption and representing the information as Linked Data.
