Survey calculations

Electronic calculation

The first electronic computers used in New Zealand government work handled calculations and data processing. The map mostly stayed on paper while the numbers went into the machine. Surveying, geodesy and engineering already generated enough calculations to occupy specialised staff and mechanical calculators, with long sequences of checking. Electronic computing could take over some of that work before organisations had a general method for storing roads, parcels or contours as digital geographic features.

A computer could reduce survey observations, solve an adjustment, transform coordinates or compare engineering alternatives without knowing that the numbers represented a road or a trig station. The result might be a coordinate list or a printed table rather than a digital map. Geography entered the computer in pieces, and the numerical pieces arrived before the graphical ones.

Computing changed the jobs of staff who had performed these calculations by hand. Lands and Survey had employed human computers to perform mathematical work long before electronic machines existed. Geodetic networks required repeated trigonometric reduction and adjustment, while engineering projects created similarly demanding calculations around gradients, earthworks and alignment. Electronic computers did not create those tasks. They changed where the arithmetic was performed and how much of it could be repeated economically.

The first government machines

New Zealand’s tidy “first computer” story is less tidy than later summaries sometimes imply. Computing-history research gives a credible case that an ICT 1201 used by the Department of Education preceded Treasury’s better-known IBM 650 by a few months. Official Public Service history gave the Treasury machine a prominent first-Public-Service role. The IBM 650 was installed in late 1960 and formally inaugurated in March 1961, which makes 1960-61 a safer boundary for the arrival of stored-program electronic computing in central government than a single trophy date.

The machines bore little resemblance to the later desktop computers on which GIS became familiar. The IBM 650 used a magnetic drum for memory and relied heavily on punched-card input and output. Programs and data were prepared before a job ran, rather than explored interactively on a screen. An operator did not drag a map, inspect a layer or click a coordinate. The computer received a carefully structured numerical problem and returned results that staff then had to interpret and use elsewhere.

Shared capacity was normal. A government agency did not necessarily own the machine doing its calculation, and the machine itself could spend the day moving between finance, statistics, science and engineering jobs. That operating model influenced early spatial computing. Survey and engineering staff had to reformulate their work so it could be expressed as inputs, calculations and outputs suitable for batch processing. The practical innovation was often as much in preparing and checking the data as in the arithmetic performed by the machine.

Cards and coordinate lists

Punched cards made that preparation visible. Each card stored a limited set of characters or numbers through holes in fixed positions. A stack could represent observations, coordinates, identifiers or program instructions, provided everybody agreed what each field meant and kept the cards in the right order. Damaged, missing or mis-sorted cards could stop a job or quietly corrupt its results.

For survey work, the attraction was straightforward. Observations that had once been copied into calculation books could be encoded, processed and checked through repeatable routines. Coordinate values could be output in consistent formats and reused in later calculations. The process was still labour intensive, because cards had to be punched and verified, programs had to be prepared, and suspicious results had to be traced back through the input. Automation reduced one kind of manual effort while creating a new need for disciplined data preparation.

The coordinate list is an important transitional object. It is digital geography in a narrow sense because positions are represented numerically and can be stored or processed electronically. It is not yet a spatial database. The computer does not necessarily know which coordinate belongs to a road, which points form a parcel, which lines touch, or which descriptive record should be linked to which geometry. Later spatial databases stored these relationships explicitly.

Geodetic adjustment

Geodesy was particularly well suited to electronic computation because its workflows were already mathematical and highly structured. A national control network combined many observations whose small inconsistencies had to be reconciled through adjustment. The work could involve large systems of equations, repeated iteration and careful propagation of corrections. Human computers and mechanical calculators could do it, but the workload increased rapidly as networks became larger and expectations of consistency increased.

NZGD49 had been completed using predominantly pre-electronic methods. Electronic computing later made it practical to maintain, recompute and transform coordinates on a larger scale. Existing trig coordinates could be loaded as data and new observations compared against the established network. Programs could repeat the mathematical operations consistently, provided the code and input were correct.

That last condition never went away. An electronic result could be wrong because the observation was wrong, the card was wrong, the program was wrong or the assumptions were wrong. The machine was fast enough to produce a large wrong answer without becoming embarrassed. Survey practice therefore retained independent checks, known control and professional judgement even as more arithmetic moved into computing.

The national coordinate framework also made computer use more useful outside geodesy. Engineering, hydrography and mapping could exchange positions because they referred to compatible control rather than entirely local numerical worlds. A coordinate became more valuable when another organisation could understand it. Long before digital files moved easily between systems, standard reference frameworks were already doing part of the interoperability work.

Engineering by batch

The Ministry of Works became one of the important homes of technical computing in the public sector. Its engineering responsibilities created problems involving routes, terrain, structures, hydrology and earthworks, all of which could require large numerical calculations. Public-sector computing histories describe substantial technical use during the 1960s and early 1970s, including road and hydro-development work. The computer was attractive where several alternatives could be calculated more systematically than a team could manage manually.

Route work illustrates the boundary between spatial calculation and GIS. An alignment can be represented by coordinates, gradients, curve parameters, terrain values and engineering constraints. Software can compare alternatives and calculate quantities without maintaining a reusable digital road network or landscape database. The problem is geographic, but the computer is solving a designed calculation rather than providing a general environment for geographic enquiry.

Retrospective accounts describe Ministry of Works projects in which computers helped generate or compare route alternatives, including work associated with the Mangaweka road-rail deviation and Clutha Valley planning. Computers supported route planning and engineering decisions before general-purpose GIS arrived.

Terrain posed a related problem. Heights could be measured at surveyed positions and software could calculate profiles, volumes or interpolated surfaces between them. By the mid-1970s New Zealand technical publications included programs for producing contours from surveyed field data. Detailed automated contouring belongs in Chapter 3 because the output was moving into computer-assisted mapping, but the underlying calculation developed from the same earlier habit: represent position and height numerically, then let the machine repeat the mathematics.

Surveying and mapping by 1966

In the 1966 Encyclopaedia of New Zealand, R. G. Dick wrote that electronic computers were being used for involved and highly mathematical surveying and mapping computations. His account describes numerical calculation as an established part of that work.

The phrase “surveying and mapping computations” covered the part of the production chain most ready for computers. Observations and control could be expressed as numbers. Transformations could be defined mathematically. Repetitive calculations could be programmed. The final map still depended on draughting, photographic reproduction, photogrammetry and other production systems described in Chapter 1.

The transition emerged through a sequence rather than a public switch-over. Calculation migrated first. A surveyor could receive electronically computed coordinates while still plotting or compiling through conventional methods. A mapping organisation could use computer-generated values while the digital linework itself remained outside the computer.

Transforming one coordinate world into another

Coordinate transformation became increasingly relevant as New Zealand modernised national mapping and measurement. Metrication changed units, while the New Zealand Map Grid provided the projection framework for the new national metric map series. W. I. Reilly’s NZMG design was published in 1973, and the new 1:50,000 NZMS260 series began appearing later in the decade. The numerical work preceded the cartographic conversion described in Chapter 3.

A projection turns positions on the curved Earth into coordinates on a flat map according to a defined mathematical relationship. Moving existing information between coordinate frameworks therefore requires more than changing a printed label. Positions have to be transformed consistently, and errors in the source control or formula propagate into the output. Electronic calculation was well suited to repeating those transformations across large numbers of points.

Survey departments already worked with coordinates, transformations and control networks, while computing departments developed repeatable numerical procedures. Bringing their work together allowed transformations to be applied systematically to large coordinate lists. These calculations made later digital conversion more practical, although they did not themselves create a digital map.

NZMG’s specialised mathematics later caused compatibility problems in international software. In the early 1970s, however, the issue was more immediate. New Zealand needed a reliable national metric coordinate framework and practical ways to convert inherited information into it. The computer reduced the arithmetic burden while survey and mapping staff remained responsible for whether the transformed result fitted the real control.

Storage without geography

Magnetic tape and disks expanded the amount of information early computers could store. A tape could hold thousands of coordinate values and still contain no explicit representation of a connected road network or parcel fabric. A program could read those values, calculate new ones and write another tape without preserving the relationships a later spatial database would expect.

A coordinate file is spatial data, but it may not be a geographic information system. A program that calculates an alignment is computer-assisted spatial analysis, but it may not support general map editing or geographic query. A plot produced from calculated values is computer-generated cartography, but it may be only an output rather than a view of a maintained spatial database.

The early systems handled demanding calculations and data-processing tasks. GIS developed through the convergence of numerical calculation, digital cartography, database design and spatial analysis. Surveyors brought coordinates and control. Engineers brought numerical terrain and route problems. Cartographers brought generalisation and graphic production. Photogrammetrists brought measured geometry from imagery. Remote-sensing researchers would bring machine-readable rasters, while land-resource teams would bring polygons and attributes. The first computerised stage solved only part of that puzzle: it made the numerical pieces easier to process.

People between the field book and the machine

Electronic computing changed the work of technical staff. Field observations still had to be made correctly, and someone had to decide how to encode them before a machine could do anything useful. Cards and tapes needed preparation and verification. Programs had to be written and tested. Results had to be checked against known values and then transferred back into survey, engineering or mapping workflows. The computer could take over the arithmetic, but it still depended on a remarkably human chain of people making sure the numbers meant what everybody thought they meant.

Some of the older occupational labels faded as the work moved. The human computer became less common because electronic machines took over much of the repeated calculation. At the same time, programmers, computer operators and technical analysts appeared alongside survey and engineering staff. The work crossed organisational boundaries because the people who understood the geographic problem were not always the people who operated the central computer.

This created a translation problem that would recur throughout GIS history. A surveyor might know what the calculation had to mean physically but not how a particular computer expected its input. A programmer might understand the machine but not recognise an impossible coordinate. Reliable work depended on both forms of knowledge reaching the same job. The later GIS specialist would often combine more of those responsibilities in one role, but the need to bridge domain knowledge and computing began much earlier.

Maps and calculation

By the early 1970s, electronic computation was no longer unusual in the technical parts of New Zealand government. Coordinates could be processed electronically, engineering alternatives could be compared, survey calculations could be repeated by program and national coordinate transformations could be handled at a scale that suited large conversion programmes. This was a substantial change from the calculation rooms inherited from the pre-war survey state.

The geography itself, however, remained mostly outside the computer. A cadastral map could supply coordinates to a calculation without its parcel fabric being stored digitally. A topographic sheet could provide terrain information without its roads and rivers existing as reusable features. An aerial photograph could support measurement while remaining a film object. The computer had become useful to mapping before the map became a computer-managed dataset.

By the late 1970s, operators were placing paper maps on digitising tables so that lines could be captured and stored as coordinates. Computer contouring, orthophoto production and analytical photogrammetry moved more of the mapping chain into digital form. The arithmetic had entered first. The machine was beginning to remember the geometry.