Database Reference
In-Depth Information
the most valuable and frequently retrieved. Tools and standards will also develop
to express GI as Linked Data more clearly, and the GeoSPARQL extensions of
SPARQL will be seen in time as a fundamental part of the query language, just as
geo extensions to SQL are now treated as just another capability of that language.
We would like to see better tools for the end user, such as souped-up browsers that
can handle geospatial queries, and we also expect to see Geospatial Linked Data
more widely used within mashup applications.
As the Linked Data Web grows, so will ontologies and reusable micro-ontologies.
These will probably not be sufficient in themselves to deal with the demands that are
placed on them to process and connect data, so it is likely that there will be develop-
ments and increased use of rule-based languages such as RIF to complement OWL.
We should not assume that Semantic Web technologies have “solved” the data
management and integration problem; it had not by any means been solved using
traditional XML and database technologies, and it still remains the most thorny issue
on the Semantic Web. However, what semantic technologies have provided is explicit
methods to aid the resolution of this problem and opened it up on the Web, and we
believe that shedding light on the dark corners of GI data integration can only make
for a happier future.
11.8 CONCLUDING THOUGHTS
Geography and the Semantic Web share a common characteristic: They are both
aids to data integration. Geography provides a means to connect information
through shared location, the Semantic Web through shared identity. Together, they
can begin to move data integration from an art and cottage industry to science and
factory. Neither geography nor the Semantic Web approach is a cure-all. Much data
has no natural geographic aspect, and some datatypes are not suitable for expres-
sion using Semantic Web technologies. But, there are always limitations with all
things, so the important thing is to understand where and when they are applicable.
Our last piece of advice is true for any new or unfamiliar topic or technology: Start
gently, start small, try to understand the underlying principles, experiment, and iterate
to build on your successes. For geography, an important starting point is to properly
understand identity and classification, that is, the nature of the things you are dealing
with. To tackle the Semantic Web, it is important to gain an understanding of the open
world assumption and then start by expressing some simple data in triple form. Once
you gain confidence, you may wish to publish your triples as Linked Data, paying par-
ticular attention to openness, reuse, and descriptions of provenance and then try creat-
ing links to other data. From there, you may wish to experiment with more detailed
description of this data by building ontologies that will further aid integration. This is
the way that we propose you get to grips with Semantic Web technologies—Linked
Data irst, ontologies later—and it is relected in the structure and order of the topic.
However, once you have got to grips with the technologies, a more appropriate
development workflow would be to start with the ontology. We strongly advise this
approach as developing the ontology will provide a systematic framework overlying
your Linked Data.
