Database Reference
In-Depth Information
by others and how they might wish to interact with it. Therefore, clever
solutions for a database that you can hide from a user may be confusing if
repeated on the Web. We expand on this in further chapters.
5. Separation between data and description: The main benefit of this is that the
description is now far more visible within an ontology; it is not locked in the
structure of the database, the query, or the application code. It also has the
consequence that others may choose to interpret your data differently from
the way you do, and likewise you may choose to interpret others' data dif-
ferently. Therefore, by applying a different ontology, the data can be inter-
preted differently. This may seem just wrong—surely you should always
use the data in the way the publisher intended? The first thing to realize is
that we often already reuse data in ways not intended by the publisher. The
Semantic Web just enables this use and the consequences of such use, to be
more explicit. Second, you may want to use the data with the same meaning
as the publisher but just need to express it differently to it with your wider
view of your world; this could be as simple as changing the vocabulary
that is used or be more complicated perhaps: changing the relationship that
one thing has with another. For example, a cadastral dataset might apply
an address to a property as a whole, its buildings and grounds, whereas
someone interested in historic buildings may wish to take that address and
apply it only to the historic building within that property. The relationships
are clearly different, but each is correct in the appropriate worldviews. And
last, as a publisher, you cannot ever plan in advance for all the different
ways in which your data might be reused.
4.5 SUMMARY
The adoption of Linked Data, and more broadly the Semantic Web, will enable
the GI community to exploit the value of GI more widely and in doing so will both
expand its use and enable it to better it within the Web as a whole. However, to
achieve this, the GI community needs to recognize that GI is not so much a hub
as glue. On the Web, with its open world assumption, there is little place for a
GI-centric view; in a minority of circumstances, this may be necessary and appro-
priate, but in most circumstances, it will present a barrier to the use of the data.
If we view GI as glue, then we begin to understand the role it more often takes
in integrating other data. Glue is rarely the focus of attention but is a component
essential for binding together so many different things. And, perhaps this is how
we should view GI: as an information glue to integrate different data. This view fits
more naturally within the Semantic Web, where the importance of explicit links
between data is so highly valued.
The Semantic Web is not the right vehicle for performing complex arithmetic
calculations and will no more replace GIS than spatial databases have. From under-
standing this limitation comes an appreciation of what can be achieved through the
use of topology and other relationships.
This chapter completes our introduction to the Semantic Web and Linked Data.
The rest of the topic now describes the technologies and their application to GI in
