Geoscience Reference
In-Depth Information
many of the models of geography also represent strongly recurring spatial patterns, for example,
distance decay in trip frequencies from an origin. The quantitative geographers of the 1960s and
1970s tried to develop precise mathematical models to describe these patterns, but maybe they
attempted too much precision and in the process lost the spatial dimension. The HPC revolution
of the mid-1990s provides an opportunity to become less precise and more general, by developing
pattern recognition tools that can build up recurring map pattern libraries of recurrent idealised
forms. Suppose you ask the question how many different spatial patterns do British cities exhibit?
Currently this question cannot be answered but at least the tools exist to allow geographers (and
others) to start to find out. Openshaw (1994e) argues that a more generalised pattern recognition
approach provides the basis for a new look at geographical information with a view to extracting
useful new knowledge from it. Turton (1997, 1999) provides some of the first computer experi-
ments. But this is only possible now as we enter the petaflop era, and it becomes feasible to apply
pattern templates to many millions of locations at many different levels of resolution. This is an area
ripe for further research.
1.8 GEOCOMPUTATION AND FUTURE GEOGRAPHERS
Much of modern human geography is now in a mega-mess and is indefensible. Most is not geo-
graphical, much is little more than storytelling, much is intensely theoretical and complex, nearly all
is non-science based, there is little use of any of the world's databases, large volumes are anecdotal
descriptions of the unique which are irrelevant to the needs of the modern world, and there is little or
no use of modern technology and no concern to meet the needs either of society or of commerce in a
fast-changing world; see Openshaw (1998a). Equally, quantitative geography is dead. It is no longer
taught in many geography departments, it has failed to become properly spatial, it cannot cope with
the needs of GIS, it emphasises statistical and mathematical tools that are simultaneously too com-
plex and too restrictive, and it fosters the impression of being an unattractive, out-of-date and old
legacy-based technology that is also out of touch with reality. Equally, there are problems with GIS,
which are just about management and capture of mappable information and have steadfastly failed
to develop beyond its original map-based origins. As a result, GIS remains surprisingly narrowly
focused, it is lacking in theory, it is largely devoid of much of the modelling and simulation relevant
to the modern world, and it is limited in what it can deliver. The time is ripe for something new that
can build on existing human and physical geographies; relates to GIS without being restrained by
it; makes use of modern informatics; is exciting, relevant and applied but not devoid of theory; has
depth but is not exclusive; and is acceptable to other sciences. It also needs to be able to invigorate
and excite new generations of geographers with a development path that links past and present to
the future and which also allows first-year students to participate. It has to be both evolutionary and
able to rapidly exploit new developments in a revolutionary manner when need be. However, if GC
is to survive and expand, then it also has to be no more difficult to the beginner than writing essays
on the life and times of a particular gendered ethnic group with distinctive sexual practices in a spe-
cific rural village! The view here is that GC has the potential to meet these objectives and is likely
to develop into a major paradigm in the new millennium.
It is argued that many areas within and outside of geography could benefit from the adoption of
a GC paradigm. Couclelis writes:
… geocomputation has the potential to influence a number of other spatial sciences, disciplines and
application areas with a spatial component, but also to be noticed beyond the walls of universities and
research centres. This potential is based on the fact that geocomputation blends well with several major
trends in contemporary society. It is obviously in tune with the computer revolution, and capitalizes
on the continuing dramatic expansion in computing power and the ubiquity of user-friendly, versatile
machines. It has a better chance than stodgy quantitative geography to attract the interest of the coming
generation of researchers. Couclelis (1998a, p. 25)
