University teaching and research

In 1985, when the University of Auckland was already teaching GIS to senior undergraduates in Geography, there was no settled New Zealand pathway for becoming a GIS practitioner. The software was specialised, useful digital data were comparatively scarce, and access to suitable computing could determine what could be taught in practice. Students were learning a field while the field was still sorting out much of its own toolkit. GIS also sat awkwardly between established subjects. It involved cartography and geography, but also databases, computer graphics, image processing, statistics and the mathematics needed to make different geographic datasets line up. A university course therefore had to teach more than a new way to draw maps. It had to decide what students needed to understand about geographic information itself.

David V. Hawke described the Auckland course in a later account. His 1992 paper on gistt, a geographic information systems teaching tool, records senior undergraduate GIS teaching from 1985. The official University of Auckland Calendar for 1986 makes the subject visible in the curriculum as Geography 20.314, “Geographic Information Systems and Image Processing”. Its description included GIS principles and the computer analysis and presentation of cartographic and remotely sensed data. The combination is characteristic of the period: GIS and image processing were close neighbours because both required students to work directly with digital representations of geography before either had become routine desktop software.

The teaching environment was mixed rather than vendor-pure. Hawke's account records the use of both public-domain and commercial software. That arrangement gave students access to GIS concepts without assuming that one package defined the field, but it also meant teaching across interfaces, data structures and limitations that were not standardised. The later gistt work grew from that educational setting. Hawke developed a teaching tool suited to the course’s practical exercises.

Auckland’s teaching connected with other academic and professional GIS work. Hawke presented “Terrain modelling and GIS” at the inaugural Spatial Information Research Centre colloquium at Otago in 1989 and a paper on the influence of data complexity on GIS performance at AURISA in Wellington in 1991. Teaching, terrain modelling and computational performance were appearing in the same academic career. The subject was already moving between classroom questions and research questions about what GIS could process efficiently.

Auckland offered senior undergraduate GIS teaching in 1985, several years before the expansion of organisational GIS in the early 1990s.

Early GIS education was constrained by access in ways that are easy to lose from later accounts. A course depended on institutional software, laboratory access and prepared datasets rather than each student installing the same package on a personal computer. Commercial GIS and image-processing systems were institutional purchases, while useful digital data often came from government or research custodians under technical and sometimes commercial conditions. Laboratory time therefore had to be shared, exercises had to match the data and software actually available, and staff had to prepare material before students could begin the spatial part of the work.

Courses combined geographic concepts with practical software use. Students had to learn enough about coordinates, spatial data structures, raster and vector information, database relationships and map output to understand why an operation worked, while also learning the commands needed to make a specialist system perform it. A failed exercise might reflect a conceptual mistake, a data-format problem, a projection mismatch, a software limitation or simple lack of access to the required file. Teaching GIS meant teaching students how to separate those possibilities.

Students learned data preparation through their practical exercises. National datasets were not yet waiting in open portals, and many New Zealand spatial records were still being digitised inside government agencies. A lecturer designing a practical exercise therefore needed data that were small enough to handle, documented well enough to teach from and legally or institutionally available for student use. Remote-sensing exercises added image files, classification and display requirements to the same problem. The close pairing of GIS and image processing in Auckland's 1986 Calendar and Canterbury's later ARC/INFO and ERDAS environment reflects a teaching world in which the boundaries between spatial database work and digital imagery were still being built in practice.

Course material also had to compensate for a field that was changing faster than a settled local textbook tradition could capture. Staff used software manuals, international literature, their own research and local datasets; Hawke also developed a dedicated teaching tool at Auckland. The first SIRC proceedings soon provided another source of material grounded in problems being discussed in New Zealand. A student could encounter terrain modelling, organisational implementation, data exchange or natural-language spatial queries through work produced by researchers they might hear at a local colloquium. Teaching and research were therefore feeding each other from the beginning.

The University of Canterbury followed a different route. A near-contemporary review by John P. Wilson and Pip Forer records Geography entering GIS teaching in 1987 with a graduate course. In 1989 the department acquired workstation ARC/INFO and ERDAS capability. The acquisition put a general vector GIS and a major raster image-processing system into the same teaching and research environment, allowing students to work with geographic databases and remotely sensed imagery rather than treating the two as unrelated specialities.

Workstation GIS expanded the exercises that could be attempted in a university laboratory. Forer and Wilson's 1993 account of Canterbury is unusually concrete: Geography's Sun SPARCstations were joined by a Roland eight-colour pen plotter, a Numonics digitiser, inkjet printers, a 24-bit scanner and Magellan Navstar 5000 GPS units. The teaching laboratory used fifteen Acorn A5000 RISC workstations as X terminals into the Sun environment, while Geography deliberately avoided PC ARC/INFO. Elsewhere on campus, departments had accumulated Atlas-Pro, Tech-Base, GRASS, Intergraph MGE and other PC-based systems and, in Forer and Wilson's words, there was “little standardization of software at the present time”. The Havelock prototype tried to turn that mixed environment into something ordinary students could use, bringing scanned aerial photography, remotely sensed imagery, vector maps, bibliographic material and analytical tools together through Acorn RISC-OS, Sun OpenWindows and the campus network. It was less a neat modern GIS lab than a collection of specialised machines and packages being persuaded to cooperate. The equipment still required technical support, system administration and staff who could keep software and datasets usable between classes, people whom surviving publications generally name less readily than lecturers and researchers.

Canterbury also demonstrates how quickly GIS escaped the boundaries of a geography course. In 1990 Richard Pascoe and John Penny published work on the interchange of data between geographic information systems and on constructing interfaces for geographic-data exchange. Pascoe and Neville Churcher then presented “Sharing geographical data” at the second SIRC colloquium in 1990 and developed the work in the New Zealand Journal of Computing the following year. By 1992 Canterbury computer scientists were presenting GIS as a set of problems of direct interest to computer science.

The papers examined spatial information and its use. It was the difficulty of moving geographic information between systems whose structures and assumptions differed. Data interchange required decisions about geometry, attributes, identifiers and the meaning of the records being transferred. Modern web services and open standards did not yet provide the later infrastructure for routine exchange. The Canterbury work records researchers treating incompatibility itself as a research problem while operational GIS users were beginning to experience the same problem in practice.

That split between Geography and Computer Science is useful for understanding how a university field forms. Geography could teach spatial analysis, cartography, remote sensing and applications. Computer Science could investigate data structures, interfaces, databases and collaborative software. Surveying and engineering brought other spatial measurement traditions. Courses, research papers and projects brought these disciplines together.

The early curriculum could easily have become little more than software training because the packages were unfamiliar and the interfaces demanding. The research record shows a wider direction. Hawke was studying terrain modelling and computational performance. Canterbury researchers were working on data interchange. At Otago, papers soon addressed natural-language access to GIS databases, semantic data models, fuzzy spatial representation and organisational GIS adoption. Universities were using the new systems while also examining where their underlying models failed or created new problems.

The terminology needs care. Much of this work can now be recognised as part of GIScience, but New Zealand researchers did not necessarily use that label for every project at the time. Their paper titles more often referred to GIS, spatial information, data modelling, terrain, information systems or particular applications. The intellectual shift is visible without retroactively renaming all of it. Researchers were asking questions about geographic information that could not be answered by learning a sequence of software commands.

The publication trail also shows the value of universities as places where awkward problems could be examined without first having to justify a national production system. A paper could test a method for exchanging geographic data, representing uncertainty or querying a spatial database before an agency had a business case for implementing the technique. A thesis could spend a year or more on a problem too narrow or speculative for an operational team. Some of that work remained academic, while other ideas moved outward through graduates, conference papers and collaborations.

Universities were also beginning to confront who spatial information was for. Vaughan R. J. Payne’s 1991 University of Otago MA thesis was titled Urupare Torohiko: Planning for the Equitable Use of Spatial Information in the Resolution of Treaty Claims. Payne examined equitable access to spatial information in Treaty claims.

The University of Otago brought researchers together across courses and departments. On 30 November and 1 December 1989 the Spatial Information Research Centre held its inaugural colloquium in Dunedin. The timing placed it only five months after the national multidisciplinary GIS conference in Wellington, but the two events served different historical roles. The Wellington meeting brought a broad professional community together at a moment of convergence. SIRC created a recurring university-centred forum in which spatial-information research could be presented, tested and revisited year after year.

The first colloquium already mixed technical and organisational questions. William Mackaness presented “Introducing GIS Into Organisations”, while David Hawke presented terrain modelling. The second colloquium in 1990 included P. J. Sallis on natural-language information retrieval from a GIS database and the Canterbury work on sharing geographic data. By the third meeting in 1991, the proceedings carried the title GIS: Realising the Potential and included work on multimedia spatial knowledge. SIRC also dealt with organisational, data and research questions beyond software demonstrations.

The fourth colloquium in 1992 illustrates how broad the research agenda had become. M. R. Anderson and George Benwell surveyed GIS use among local authorities in New Zealand and Victoria. P. R. Zwart wrote about mature GIS and the idea of “taking the G out of GIS”, while Canterbury computer scientists outlined GIS problems of interest to their discipline. These topics linked organisational adoption, information-system integration and technical research. The university forum was watching GIS spread into practice while simultaneously asking what the technology was becoming.

By 1993 and 1994, SIRC papers included semantic spatial data modelling, fuzzy representation, hazards information systems, wildlife decision support, connectionist approaches to spatial information, GPS data capture and repeated surveys of local-government GIS use. The Dunedin Pilot Hazards Information System brought University of Otago researchers together with geological scientists and local-government information requirements. In 1994 the project was reported as a GIS trial using natural-hazard, topographic and cadastral information. The university was participating directly in the field, with research projects crossing institutional boundaries.

The SIRC proceedings also reveal the advantage of recurrence. A single conference can preserve a snapshot, but an annual series records changes in questions. The local-government surveys could be repeated. A proposed hazards system could return as a trial. Methods first presented as conceptual work could later appear in applied projects. By the seventh colloquium in 1995, held jointly with AURISA in Palmerston North, papers were addressing environmental management and natural-language techniques for railway safety reporting. The forum had become a continuing part of the Australasian spatial-information research network.

The original annual-colloquium sequence can now be traced from 1989 to its twentieth event in 2008. That later history belongs elsewhere in the book, and the continuity confirms that the early meetings were more than temporary experiments. The inaugural SIRC NZ GIS and Remote Sensing Research Conference in 2013 began a separate series.

A research field also needs places where work can be recorded. SIRC proceedings became one such archive, but they sat beside journals and professional publications rather than replacing them. Hawke's teaching work appeared in the International Journal of Geographical Information Systems. Canterbury data-interchange research appeared in the New Zealand Geographer, the New Zealand Journal of Computing and an international GIS journal. The same spatial problem could therefore be read by geographers, computer scientists and GIS specialists through different publication channels.

New Zealand's specialist cartographic literature was changing at the same time. The New Zealand Cartographic Journal ran to 1990, and its successor from 1991 carried the title New Zealand Cartography and Geographic Information Systems. The journal brought cartographic and GIS work into the same publication. Universities were teaching and researching GIS while cartographers were debating how computer mapping and GIS affected established practice.

The resulting literature was fragmented by modern standards. A useful idea might first appear in a colloquium paper, return in a thesis, then be developed in a journal article or an operational project. SIRC's recurring proceedings are valuable because they preserve early stages that conventional journal databases often miss. Students and researchers used conference volumes, software documentation and journals from several disciplines.

Several university disciplines adopted GIS. Otago's surveying tradition already dealt with coordinate systems, measurement, photogrammetry and the legal and technical organisation of land information. Canterbury's computer scientists were examining exchange formats and interfaces while geographers worked with GIS and remote sensing. Environmental and geological researchers brought problems defined by terrain, hazards and natural resources. Planning supplied questions about land use, administration and access to information. The emerging field was held together by geographic data rather than by one departmental lineage.

Surveying and GIS research overlapped at Otago. SIRC papers involved spatial measurement, GPS, hazards, databases and information systems, and the university's research community crossed departmental boundaries. That mixture helps explain why the centre used the broader term spatial information rather than defining its remit only around one GIS package or one academic discipline.

This multidisciplinary setting also affected what counted as research. A computer scientist could treat GIS as an interface and data-structure problem. A geographer could examine terrain or local-government adoption. A surveyor could focus on positioning and measured spatial data. A planner could ask how information should be used in decisions. Those questions shared enough concepts and tools to create a field while remaining different enough that no single department could easily own it.

George Benwell appears repeatedly in the early Otago record, particularly where GIS met organisational and environmental questions. His work with other researchers turned the adoption of GIS into a subject for empirical study. That research was useful because the rapid spread of GIS during the early 1990s created claims about benefits faster than organisations could always demonstrate them. Universities could compare users, examine implementation patterns and ask why some systems became embedded while others remained specialist tools.

The local-government work developed through several stages. A 1992 SIRC paper surveyed GIS usage among local authorities in New Zealand and Victoria. A 1994 paper compared the New Zealand situation in July 1993 with February 1992. Andrew Marr and George Benwell then extended the work through a 1995 study of local-government GIS since 1989. Marr's 1996 MSc thesis, Geographic Information Systems Maturity in New Zealand Local Government, provided a 126-page research treatment of the subject.

Chapter 23 describes local-government adoption in more detail. The university history is in the method and the feedback loop. Practitioners were installing systems; researchers were surveying what had happened; the findings were presented back through SIRC and other forums; postgraduate work then developed concepts such as GIS maturity and organisational integration. University researchers examined how organisations implemented GIS.

That connection also helped produce practitioners who understood GIS as an information-system problem. Students encountering the field through postgraduate work saw data quality, organisational structure, maintenance and user requirements alongside spatial analysis. A thesis on maturity or a project on hazards information required engagement with organisations outside the university. The line between education and professional practice was therefore permeable well before formal geospatial degree programmes became common.

Theses are particularly useful in the history of a young technical field because they preserve work that may never become a journal article. They can record data structures, software limitations, implementation choices and failed approaches in more detail than a short paper. Colley's 1979 Canterbury thesis is an earlier example, but Chapter 3 owns that MAPPAK story. By the 1990s, theses and postgraduate projects were appearing within an established GIS research community.

Payne's 1991 Otago thesis and Marr's 1996 MSc illustrate how wide that environment had become. One addressed equitable use of spatial information in Treaty claims; the other studied GIS maturity in local government. Neither can be reduced to learning a package. Both treated spatial information as part of larger institutional relationships. Other university work during the period connected GIS with hazards, environmental management, wildlife, terrain modelling and data exchange.

The postgraduate system also moved knowledge between organisations. Research problems came from councils, science agencies, government records and field projects. Students and staff presented results at SIRC, AURISA and specialist meetings. Graduates then entered a labour market in which councils, government departments, consultancies and science organisations were expanding GIS capability. The universities did not create that demand, but they increasingly supplied people able to work with geographic databases as well as maps.

Technical staff maintained the teaching environment. Workstation laboratories needed machines configured, accounts managed, software installed, data copied, plots produced and practical sessions supported. Course demonstrators and technical staff often appear poorly, if at all, in journal papers and conference proceedings.

By the middle of the 1990s, New Zealand university GIS had several characteristics of a durable field. Formal teaching was documented at Auckland and Canterbury. Otago had a recurring spatial-information research forum with a growing proceedings series. Computer scientists were publishing on geographic data exchange and system design. Postgraduate research was examining organisational adoption, Treaty-related spatial information, hazards and environmental applications.

The field was also connected outward. SIRC papers included researchers from universities, Crown science organisations and practical government contexts. AURISA provided another route between academic and professional work. GIS education was therefore not developing behind the university gate while practitioners built a separate industry. The same people, papers and research questions moved between laboratories, departments, conferences and operational organisations.

Software remained necessary, but it was becoming a less satisfactory way to define expertise. A student trained only to operate one package would still confront incompatible data, uncertain source quality, database design, organisational maintenance and spatial-analysis questions that the interface could not solve. The early research record shows universities already working on those harder problems. Teaching had begun with scarce access to specialised systems; research was turning GIS into a subject that could critique and extend those systems.

By 1996 there was still no single national GIS curriculum and no reason to pretend every university had developed at the same pace. First-course dates remain unresolved for Otago, Waikato, Victoria, Massey and Lincoln, and the surviving record is stronger for published researchers than for technical teaching staff. The documented pattern is sufficient without filling those gaps. New Zealand had university GIS courses, specialist laboratories, postgraduate research, an annual spatial-information colloquium and regular exchange with practitioners.

That exchange extended beyond the universities. Vendors, consultancies, user groups and professional networks were creating a wider GIS community at the same time that universities were producing research and graduates. By the mid-1990s GIS in New Zealand was no longer held together only by a few departmental systems or isolated technical projects. It had acquired places where it could be taught, criticised, researched and passed to the next group of practitioners.