The Dictionary of Human Geography

to map forest cover in order to recommend
locations for new growth; Pratt worked for
the Canadian Ministry of Agriculture, which
wanted to compile land-use maps for the entire
country, maps that would describe multiple
characteristics including agriculture, for-
estry, wildlife, recreation areas and census di-
visions. Tomlinson suggested that they pioneer
a computerized system in which land-use zones
were digitally encoded so that they could be
overlaid with other relevant layers such as
urban/rural areas, soil type and geology. This
happenstance meeting led, in 1964, to the Can-
ada Geographical Information System
(CGIS). The name of the system was bestowed
by a member of the Canadian Parliament.
There were parallel developments in
Europe. Tom Waugh, for example, developed
an early GIS system with the acronym
GIMMS. It was avector-based GIS system
with sophisticated analysis, and was eventually
used in twenty-three countries (Rhind, 1998).
GIMMS preceded ESRI (see below) in devel-
oping a commercial GIS and was relatively
sophisticated for the 1970s and 1980s – includ-
ing cartographic options and batch processing.
In the USA, the Harvard Graphics Laboratory
was a tinderbox of the GIS revolution. Re-
search at the laboratory established an efficient
method for computerized overlay using poly-
gon (vector) boundaries. The laboratory was
populated by a host of researchers who had a
profound influence on the development of cur-
rent GIS, including Nicholas Chrisman and
Tom Poiker. A diaspora of researchers from
the Harvard Laboratory in the 1970s contrib-
uted to the dissemination of GIS, especially into
the private sector. Scott Morehouse, a junior
member, left in 1981 to work for a company in
California called Environmental Systems Re-
search Institute (ESRI), where he re-developed
the algorithm for vector overlay which became
a cornerstone of the program ARC/INFO. This
dispersion of ideas from the Harvard Laboratory
was the beginning of one GIS identity: that
linked to software packages, hardware systems
and technology in general (Chrisman, 1988).
Institutional and governmental support for
GIS was also a major impetus for its growth
and adoption from the 1970s onward. In the
UK, four multidisciplinary Regional Research
Laboratories (RRLs) were designated by the
Economic and Social Research Council. They
were designed to facilitate primary functions
of GIS, including spatial data management,
software development, spatial analysis and
training of GIS researchers (Masser, 1988). In
the USA, the National Center for Geographic

Information Analysis was funded by the

National Science Foundation (NSF). Three

US universities with GIS expertise were

chosen as primary research centres. Their

role was to facilitate understanding and imple-

mentation of geospatial methodologies and

develop university adoption of these tech-

niques. The NCGIA also played an important

role in hosting and responding toepistemo-

logicaland pragmatic critiques of the tech-

nology (Pickles, 1995b; Curry, 1998).

The development of GIS, however, is not

rooted solely in computer laboratories and

universities in the latter part of the twentieth

century. It is also an outgrowth of attempts to

automate calculation in the nineteenth cen-

tury reflected in efforts, for example, to code

population data for the UScensusin 1890

(Foresman, 1998). Pre-eminent GIS scholar

Michael Goodchild makes the point that GIS

was developed during a period when informa-

tion was increasingly being translated into

digital terms and disseminated widely (Good-

child, 1995). If geographers had not explored

the possibilities of digital manipulation of spa-

tial data, other disciplines would have initiated

the process. As it is, many roots of GIS are in

disciplines other thangeographyincluding

landscape architecture andsurveying. Like

all technologies, GIS is an outcome of both

social and technological developments. ns

Suggested reading

DeMers (2000); Longley, Goodchild, Maguire

and Rhind (1999); Schuurman (2004).

geographical analysis machine (GAM) An

example of automated spatial data analysis

catalysed by three factors: the growing avail-

ability of digital data with point (x,y)geo-

coding; a move from statistical techniques

‘smoothing over’ geographical variation to

local statisticsrevealing geographical pat-

terns in data; and increased computational

power to guide where to look. GAM passes a

moving window of fixed radius (or population

count) across a studyregion, repeatedly test-

ing for unusual clusters of a particular feature.

Successfully used to study the clustering of

cancers, GAM and its primary architect –

Stan Openshaw – inspired much of the

research ingeocomputation. rh

Suggested reading

Openshaw (1998).

geographical explanation machine (GEM)

Whilst thegeographical analysis machine

Gregory / The Dictionary of Human Geography 9781405132879_4_G Final Proof page 281 2.4.2009 6:30pm

GEOGRAPHICAL EXPLANATION MACHINE (GEM)