The clandestine computer on The Terrace
It is generally considered that the first stored-program electronic computer in New Zealand was an IBM 650 installed in late 1960 by the New Zealand Treasury and inaugurated in March 1961. However, there is good reason to believe that, in fact, the first such machine was an ICT 1201, installed slightly earlier for the Department of Education (not a Ministry at that time) and that this machine was already processing the fortnightly teachers’ payroll by August or September 1960. Unlike the Treasury IBM machine, this computer received no publicity and was quietly installed in ICT’s New Zealand headquarters in Shell House (now Transpower House), at 96 The Terrace, Wellington. Later (probably in 1962), it may have been moved to a Department of Education building. Its next stop, in 1963, was at the Motor Specialties headquarters building on Anzac Avenue in Auckland. In 1966 it was donated to finish its working life at the Auckland Technical Institute (now Auckland University of Technology). There is no clear paper trail for this history, especially for the machine’s date of arrival in New Zealand, but the available evidence is laid out in the article, ‘The First Computer in New Zealand’ (Carpenter 2020). For the sources of most of the following story, please see the references in that article.
Why was this computer such a secret? To understand this, we have to put ourselves back into the minds of Wellington’s civil servants in the late 1950s, or indeed considerably earlier. It was in 1920 that data processing by machine became a reality in the New Zealand government, when Malcolm Fraser, the Government Statistician, returned from the British Empire Statistical Conference in London with a strong recommendation that the 1921 census data should be processed using Powers punched card equipment. By contrast, for the same purpose, Australia had chosen Hollerith equipment, as Fraser discovered during a stop in Melbourne near the beginning of his voyage to London. He was delayed by a dockers’ strike in Cape Town, so he was able to have discussions with his South African colleagues, who were also committed to Hollerith. On his way home from London, he spent some time on the East Coast of the USA investigating the usage of Powers and Hollerith equipment for statistical analysis, and his report to the government includes a detailed analysis leading to a final recommendation in favour of Powers. Not only did he slightly favour this equipment technically, but also he was concerned that Hollerith equipment could only be leased, not purchased, in New Zealand. As a result of his report, Powers punched card equipment quickly entered into government service.
As an aside, the 1916 New Zealand census was therefore the last one whose results were compiled manually. The census report1 notes condescendingly that:
1 Fraser, Malcolm, ‘Report on the results of a census of the population of the Dominion of New Zealand taken for the night of the 15th October, 1916,’ www3.stats.govt.nz/historic_publications/1916-census/ [retrieved 20 March 2021]
A feature of the compilation of the ‘Results of the 1916 Census’ was the employment of female clerks for the bulk of the work, a departure which was necessitated by war conditions, and which was found to work well enough in practice.
A noticeable feature of the punched card era that started in 1921 was that the people who operated the noisy, demanding machinery were almost all women, supervised by men. By 1946, the starting weekly salary for a 16-year-old Powers-Samas or Hollerith machine operator was £2/1/6 ($4.15).
To clarify the terminology, note that ‘Hollerith’ in the 1920s referred to the card format standardised by the Computing-Tabulating-Recording Company (CTR) in the USA, renamed in 1924 as IBM. The corresponding equipment was supplied in Britain and the Dominions by British Tabulating Machinery Ltd (BTM), which had operated since 1908 under an exclusive but restrictive contract with IBM and its predecessors. ‘Powers’ referred to the format and equipment standardised by the Powers company in the USA, later absorbed into Remington Rand. From the 1930s to the 1950s, the British company known as ‘Powers-Samas’ exclusively supplied such equipment in New Zealand.
As well as the Government Statistician, the New Zealand Railway Department also started using punched card equipment for accounting and statistical purposes in the early 1920s and business usage of punched cards must have begun soon afterwards. The British Powers-Samas company established itself locally, eventually having offices and technicians in all the main centres. Somewhat surprisingly, its main rival in the British market, BTM, colloquially known as ‘Hollerith’ in Australasia, established itself in Australia but not in New Zealand. For this and other reasons, government data processing became more and more dependent on Powers-Samas equipment. By the end of World War II, most government and private users of punched cards were Powers-Samas customers. There was some usage of Hollerith equipment, for example by the Reserve Bank but this was still supported from Australia throughout the 1950s.
This seems very odd in hindsight, because from the earliest days of electronic computers, the card format that dominated was the 80-column format developed by IBM and known in the UK and Australasia as ‘Hollerith’. As noted above, this format and the machines that processed it were supplied and supported only by BTM, even if they had no office in New Zealand. However, in the period from 1920 until the late 1950s, choosing the Powers-Samas equipment was a perfectly rational decision. In government, it was enforced by Treasury, which was fully committed to Power-Samas and had to approve all major expenditure by other departments. Reading between the lines in the archives, it is clear that Hollerith purchases were discouraged by wielding the most powerful tools at Treasury’s disposal: budget approvals and import licences. This situation changed quite suddenly in early 1959, as we shall see.
For most of the 1950s, the distinction that we now consider essential between an electronic calculator and a stored-program computer was not so clear. Both BTM and Powers-Samas sold electronic calculators with small magnetic drum memories, namely the BTM 555 and the Powers-Samas Programme Controlled Computer (PCC). In both cases the ‘programme’ was external, set on switches in the case of the PCC or a plug board in the case of the 555. For the British market, an important feature of these machines was their ability to calculate in pounds, shillings and pence, which of course gave them an advantage in New Zealand too. Records of the sales of these devices in New Zealand are not available but components of a BTM 555 are to be seen in Toitu Otago Settlers Museum, Dunedin. However, despite their speed advantage over electro-mechanical card processing equipment, their programming remained a clumsy manual process, so they were hardly more flexible than a plug-board controlled mechanical card sorter.
By the late 1950s, the advantages of internally stored programs were clear and the modern computing era began. In the UK, Powers-Samas was completely unprepared for the computer era. Their rival BTM had been freed from its restrictive agreement with IBM in 1949. BTM introduced their electronic calculator several years before Powers-Samas and marketed its stored-program electronic computer, the HEC4, also known as the BTM 1201, from late 1956. Almost inevitably, given the rapid rise of IBM as a direct competitor after 1949, Powers-Samas was merged into BTM to form International Computers and Tabulators Ltd (ICT) in March 1959, and the HEC4 was rebranded for the second time as the ICT 1201. With Powers-Samas merged into its Hollerith rival, the New Zealand Treasury necessarily abandoned its policy of preferring Powers-Samas equipment and looked to IBM instead.
The two contenders
What were these two computers, the Department of Education’s ICT 1201 and the Treasury’s IBM 650? Both were very much first-generation computers, based on thermionic valve (vacuum tube) electronics, with a magnetic drum as their main memory. Both were therefore room-sized and ran hot. There were panel switches but of course no screen or keyboard. Input was from a card reader, and output was either freshly punched cards or upper-case printing. They could therefore readily be integrated into an existing punched card operation but only if the card formats were compatible. Both the IBM and the ICT used Hollerith 80-column cards incompatible with existing Powers-Samas machinery. This effectively left all government departments with no choice but to progressively abandon Powers-Samas cards and equipment as the computer age began.
BTM publicity photo
Both computers were hard to program by modern standards. To a first approximation, all programs for the ICT 1201 (Figure 1) were written in machine code on specialised coding sheets; even a fully fledged assembly language was not available. Coding was particularly tricky because the design required ‘optimum programming’, whereby the flow of code was matched to the rotation speed of the drum memory. In contrast, the IBM 650 (Figure 2) had a Symbolic Optimal Assembly Program (SOAP), among other software. Although IBM did develop Report Program Generator (RPG) in 1959, allowing convenient mapping of punched card operations into computer code, it was for the IBM 1401, not the 650. ICT had nothing of that kind for the 1201. In any case, the drum memories were tiny by today’s standards (as little as 4 kilobytes in the case of the ICT 1201), so extreme economy of programming was required. However, the 650 was an extremely successful business and scientific computer by 1960, built in the US as fast as one per day, so programming it for Treasury’s needs would have been fairly straightforward. This could probably not be said of Education’s ICT 1201.
Competition for government business
In New Zealand, the Powers-Samas registered office in the Pharmacy Building at 59 Cambridge Terrace (Figure 3), Wellington, rebranded itself as ICT at the end of June 1959. The move to Shell House, much closer to all government departments, followed soon after. ICT later evolved into ICL and moved to 126 The Terrace, until acquired by Fujitsu, now at 141 The Terrace. Thus, in early 1959 all Powers-Samas employees woke up one morning as ICT employees, which must have been a strange experience since they had been trained to think of Hollerith as the enemy. Now IBM was the enemy. As early as 1957 the New Zealand Treasury had signed a contract to lease an IBM 650, but when ICT started operations in Wellington this machine had not been delivered. The race between IBM and ICT was on.
Treasury documents in the National Archive (Archives New Zealand, n.d.) show that within a few years of the first IBM delivery in late 1960 government data processing equipment was dominated by IBM imports, in stark contrast to its domination by Powers-Samas since the 1920s. The file of requests for computer import licences between 1965 and 1975 contains a long sequence of letters of application from IBM in almost the same words except for the name of IBM’s customer in each case. (A notable exception was the import of Burroughs B6700 machines for five universities, negotiated in 1971.) From having no presence in the NZ market, IBM went from zero to hero within a few years. This was a delayed but direct result of the termination of their agreement with BTM in late 1949, which at the same time as freeing BTM to develop their own computers, released IBM to compete with them for business in Commonwealth countries. However, it was not until 1957 that IBM Australia secured that crucial first contract to lease an IBM 650 to the NZ Treasury.
Although many factors were at play, including various government requirements for statistical analysis and the DSIR’s need for much more calculation power than mechanical or electro-mechanical machines could provide, the main concern of both Treasury and the Department of Education by the mid-1950s was the mundane business of payrolls. Treasury had to calculate salaries and print cheques for 20,000 public servants every fortnight, and that number was steadily increasing. Education had to pay some 13,000 teachers around the country, and due to the post-war baby boom, this number was rapidly growing. This was not handled by Treasury, although of course the money came from them. On the contrary, New Zealand then had twelve local education boards, each managing the payroll for its area. Teachers were paid monthly, but their union (known today as NZEI Te Riu Roa) was pressing urgently for fortnightly payments. It was common ground between Education and Treasury that this required centralisation of the work in Wellington, and that Education would need some kind of modern machinery to achieve this. As early as August 1954, Treasury agreed that Education could spend a modest amount of money on Powers-Samas punched card equipment for this purpose, preferring this offer to one from Hollerith since the latter had no local support office.
The Treasury files are largely silent on this question from then on, unless a skilled archivist can discover more. But as noted above, in 1957 Treasury effectively deserted Powers-Samas by committing to an IBM 650. And in the end, whatever had happened about the 1954 proposal for Powers-Samas equipment, the Department of Education eventually picked an ICT 1201. The quirk was, however, that the people who proposed the ICT machine had been working for Powers-Samas until the formation of ICT in early 1959.
The annual budget summaries reported to Parliament by the Department of Education give us some more hints about what happened. There is no trace in the relevant years of capital expenditure on a computer. Before 1957 even punched card equipment did not figure. From then onwards, payments for ‘Hire of punch card machines’, ‘Duty & tax on accounting machines’ and an unexplained increase in ‘Office equipment’ grew rapidly to approximately £33,000 per year by 1960–61. The price of an ICT 1201 being £30,000 to £40,000 according to various sources, that amount of operational expenditure would have been consistent with renting the ICT machine, conveniently housed in ICT’s new offices on The Terrace, close to most government buildings (Figure 4).
Another aspect of the timeline is that teachers received their eagerly awaited fortnightly salaries starting in April 1959, clearly based on some kind of central automated process. Just over a year later, in September 1960, their union journal wrote:
It is anomalous that when teachers’ pay and deductions are processed by electronic bookkeeping, the methods of recruiting teachers have not changed much since the horse-and-buggy days.
This is the nearest we have to a smoking gun. Although both Powers-Samas and BTM had produced electronic calculators there is no evidence that either of them were present in Wellington then, so the only electronic candidate is the ICT 1201 on The Terrace. One plausible speculation is that whatever punched-card solution was implemented in April 1959 was quickly found too tricky or too slow. In any case, evidence from the UK indicates that the 1201 for New Zealand was ordered in mid-1959. In all probability it was up and running the teachers’ payroll at the latest in September of the following year.
The late Professor Robert W Doran started investigating early ICT computers in New Zealand but was unable to complete his work (Doran 2012). He had directly or indirectly contacted former ICT staff Jack Morgan and Tony Watts. The latter stated that the ICT 1201 was initially installed on the ground floor of Shell House and later moved to the Department of Education, possibly in Hobson Street in Thorndon, Wellington and probably in 1962. We know from other sources that Education acquired a transistor-based ICT 1301 in 1962; Watts stated that this machine in turn went into Shell House. We also know that the 1201 was sold on to Motor Specialties in Auckland in 1963, so these dates are consistent.
Another important piece of anecdotal evidence comes from Bruce McMillan, originally employed and trained by Powers-Samas in Dunedin, who later installed the famous ICT 1301 at Cadbury’s, and more recently reinstalled it in Toitu Otago Settlers Museum. He recalls that when he was in Wellington for ICT training courses, the staff there were furious about IBM’s claim to be first, because they had personally installed the first computer in the country, months before IBM. Of course, it is no surprise that IBM trumpeted the delivery of the ‘first’ computer in New Zealand, and that Treasury was happy to be seen at the forefront of technology in 1961.
The intriguing question is why the Department of Education and ICT did not publicise their achievement the previous year. We are reduced to speculation. At that time, when Treasury still had its thumb firmly on the supply chain by means of import licensing, and the Department of Education seems to have hidden its computer adventure in the punched card section of its operational expenses, perhaps the Director of Education felt that publicity was the last thing he needed. We might also suspect that ICT, exhibiting British reserve, was less eager for publicity than IBM, which was clearly trying to achieve a dominant position in the emerging computer market. IBM has of course always been highly effective at public relations; they even got their eventual donation of Treasury’s 650 to Auckland’s Museum of Transport and Technology (MOTAT) into the newspapers in March 1968. This machine had been a small gold mine for IBM, quite apart from its public relations value; rental costs for an IBM 650 were about £70,000 per year, based on US prices and contemporary exchange rates.
The historical record
A quarter of a century after computers first arrived in New Zealand, when the New Zealand Computer Society celebrated its silver jubilee, some of the history was written down for the first time, with the Treasury machine heading the list. For whatever reasons, those who knew about the Department of Education machine kept quiet. There the story would have ended, if it were not for one of the country’s first computer programmers, Ruth Engleback. Originally trained as a programmer by BTM in England in 1954, she left the company in 1959, shortly after hearing about the ICT 1201 order from New Zealand. In 1963, now living in Auckland, she was brought in to work on that very machine after it was re-installed by Motor Specialties. Finally in 1966, she was again brought in to assist Auckland Institute of Technology after they were given the machine. In 2012, annoyed by the ongoing story that the IBM was first, she alerted Prof. Doran to the truth of the matter. He published a blog entry, but was unable to complete the investigation, which fell to the present author.
We can now conclude with a complete timeline of the relevant events:
1921 New Zealand government adopts Powers-Samas punched card processing equipment.
1954 IBM ships the first IBM 650 computer in the USA.
1954 Treasury approves Department of Education adoption of Powers-Samas equipment for centralised teachers’ payroll.
1956 BTM ships the first HEC4 computer in the UK, later rebranded as the BTM 1201.
1957 Treasury commits to IBM 650 for public servants’ payroll (then prepared using Powers-Samas punched card equipment).
March 1959 Merger of Powers-Samas and BTM as ICT Ltd in the UK. The BTM 1201 is rebranded as the ICT 1201.
April 1959 Department of Education starts processing teacher’s fortnightly payroll in Wellington (presumably with punched card equipment).
June 1959 Powers-Samas New Zealand starts trading as ICT. Soon moves into Shell House on The Terrace.
Unknown date in mid-1959
ICT in the UK receives an order from New Zealand for an ICT 1201.
Unknown date in 1960
An ICT 1201 is installed in Shell House.
September 1960 Teachers’ salaries are now processed electronically, presumably on the ICT 1201.
November 1960 Treasury’s IBM 650 is installed.
March 1961 Treasury’s IBM 650 is formally inaugurated. Its most important application is the public servants’ payroll.
1962 The Department of Education acquires an ICT 1301 to replace the 1201.
1963 The ICT 1201 is sold on to Motor Specialties in Auckland.
1966 The ICT 1201 is donated by Motor Specialties to the Auckland Institute of Technology.
1968 The Treasury IBM 650 is donated by IBM to MOTAT.
Unknown date The ICT 1201 is also believed to have been donated to MOTAT.
1985 After 25 years of computing in NZ, it is generally believed that the IBM 650 was the first computer in the country.
2012 Ruth Engleback alerted Prof. Bob Doran to the ICT 1201 story, and he published a blog entry.
2020 An article on ‘The First Computer in New Zealand’ was published.
IBM Distinguished Engineer working on Internet standards (1997–2007). Earlier he led the networking group at CERN, the European Laboratory for Particle Physics (1985–96). He taught at Massey University from 1974 to 1976. He chaired the Internet Engineering Task Force (2005–7), the board of the Internet Society (2000–2) and the Internet Architecture Board (1995–2000). He holds a first degree in physics and a PhD in computer science.is an Honorary Professor of Computer Science at the University of Auckland, where he taught from 2007 to 2012. His research interests are in Internet protocols, especially the infrastructure layers, as well as computing history. He was an
Carpenter, BE (2020). The first computer in New Zealand. IEEE Annals of the History of Computing, 42(2), 33–41
ADRK 17391 T1 R15421250 Archives New Zealand Te Rua o te Kāwanatanga, Wellington
Doran, RW (2012) ‘The First ICT Computers in NZ’, incomplete notes