School of Anthropology
Geography and Environmental Studies
The University of Melbourne, Australia
The apparently imminent takeover of GISystems by GIScience, is probably a mirage. The reasons might include the diversionary interests and inhibitions found within all types of GIS teaching departments and conservatism in favour of conventional approaches. This article argues that GIScience is probably not the future face of GIS practice, or even of GIS education
In their popular GIS book, Longley and his colleagues (2002) distinguish between GISystems and GIScience. They then imply that the latter will replace the former as the dominant emphasis within GIS. Although this seems logical enough, this article will argue that GIScience, at least in the form envisaged by Longley et al, is probably NOT the future face of GIS practice, or even of GIS education.
The author has little empirical data to back up such an argument. It is developed here mostly on logical grounds. It is the author’s personal reflection from his years as a strategic planner within local and regional government, industry, commerce, consulting and, finally, academia, where he has taught GIS and environmental decision-making. This argument is still a convenient launching pad for presenting a brief sketch of GIS education within the author’s home country – Australia.
GI-systems versus GI-Science
Longley et al basically contrasts the focus of traditional GIS on technicalities and applications, GISystems, with the “science” of testing the validity of GIS’s underlying mechanisms, GIScience. For example, a GI Systems Analyst might study how to code, implement and interpret the available algorithms for interpolating points-based data across a continuous surface. By contrast, a GI Scientist would examine the validity of each alternative algorithm with the aim of determining which one is the most accurate. That is, whereas GI Systems Analysts apply techniques, GI Scientists try to verify such techniques, and it is the latter activity that will surely carry the future.
There is nothing new about wanting one’s vocation to be more “scientific” and so command the respect and prestige that is enjoyed by its cousin disciplines. Indeed, given GIS’s infiltration into an ever-expanding range of application areas, one might argue that its “scientization” is urgently required in order to reassure clients about its validity. With so much investment hinging on the results of GIS analyses, surely the users of GIS deserve to know how worthwhile GIS methods are.
Yet the same argument was put at least thirty years ago (Wyatt, 1972) within the discipline of market research. With so much marketing investment based on the results of market research, surely market research’s clients will demand that its methods be scientifically verified. But this does not seem to have happened. For long, competition within market research is vicious – tender undercutting is endemic and overselling of “findings” is rampant. The author would like to think that the situation has improved since he last looked at it, but this seems doubtful.
Now, it so happens that many of the new applications of British GIS are actually a cartographic version of market research. The rich demographic data that is available from small UK Census Collectors Districts has proved to be a boon for those seeking to build area-based “profiles” of different sorts of consumers. Such consumer profiles appear to be, on the face of it, an answer to any strategic marketer’s prayers. They open the way towards truly incisive, targeted marketing.
But there seems to have been almost no testing of the validity of the different methods that can conceivably be used for generating such area-based consumer profiles, just as there was none, decades ago, when market researchers first developed consumer profiles. Hence, both GIS and market research are proving that the presence of large amounts of money does not automatically pre-ordain that scientific testing will take place.
Besides, there are other forces acting against the emergence of genuine, scientifically rigorous GIScience. Longley et al mention some of these, but they fail to fully address any of them. For instance, the nature of education is changing markedly, with students demanding different types of delivery and different justifications for the material being presented to them, and the resulting commodification of education could well compromise “pure” research and scientific testing. Also, one company’s commercial monopolization of most of the technology on which the GIS discipline is based does not bode well for genuinely independent inquiry into alternative methods and algorithms.