And Computer Users Richard Gillespie

Collecting computers —and computer users Richard Gillespie

A door seems an odd object to the control panel. After a tantalising Among the world’s first include in a computing collection. pause, the computer began printing a computers Yet the wooden-framed glass door programmed message: The door that is now preserved in now proudly displayed in the School the collection thus symbolises the of Computing and Information Mr Vice Chancellor. university’s entry into computing, Systems (pictured below) represents Thank you for declaring me open. both for research and teaching. Earlier the beginning of more than six I can add, subtract and multiply; that year, physics researchers had been decades of computing at the solve linear and differential observed emerging from the lab after University of Melbourne. equations; play a mediocre game seeing the computer being put through On 14 June 1956, a group of of chess and also some music.1 its first trials in Melbourne, excited dignitaries and nervous staff passed by the prospect of using CSIRAC to through the door of the Computation Because CSIRAC printed at the speed up their calculations. The first Laboratory, located in what is now rate of only four or five characters course in programming CSIRAC called the Old Physics Building on per second, the assembled guests had been run in February 1956, the Parkville campus. Sir Ian Clunies- had to wait about a minute for the attracting some 30 participants from Ross, chairman of CSIRO, formally message to be completed. But they university departments, research transferred the CSIR Mark 1 had no doubt that the event marked a organisations, government and computer to the university; vice- significant moment in the university’s business. Determining the foundation chancellor George Paton then history. The newspapers reported date of a computer science department named the re-assembled computer enthusiastically about the university’s is a contentious and competitive ‘CSIRAC’ and pressed a button on ‘electronic brain’.2 endeavour, but when the Computation Laboratory at the University of Melbourne was formally established as a separate department in October 1955, it was one of only a handful in the world. CSIRAC was by then a computer with an impressive history. It had been designed and built in Sydney by British physicist Trevor Pearcey and electrical engineer Maston Beard, as a research project of the Division of Radiophysics of the Council of Scientific and Industrial Research. The CSIR Mark 1, as the computer was called in Sydney, ran its first test

12 University of Melbourne Collections, issue 22, June 2018 Opposite: Door of the University of Melbourne Below: Jurij Semkiw at the control console Computation Laboratory, 1955. CSIRAC054, of CSIRAC at the University of Melbourne, School of Computing and Information Systems late 1950s. 2017.0071.00157, University of Collection, University of Melbourne. Melbourne Photograph Collection 1857–1990, University of Melbourne Archives.

program around November 1949, making it the fourth electronic digital computer in the world. The Mark 1 was used extensively for computations in radio-astronomy and cloud physics, the two main research areas of the Division of Radiophysics, as well as by other researchers and organisations, such as the Snowy Mountains Authority. In 1951 it was even used to play some simple tunes—now recognised as being the world’s first computer music.3 By 1954, however, CSIR radiophysics chief Edward Bowen had decided to terminate the computer project, in part to focus his division’s limited funds on radio- astronomy and cloud physics, but also to pressure the government tables. Betty Laby in the Statistics computer. Extolling the powers of and universities into picking Department managed a team of the machine, Martin stressed to the up responsibility for electronic women ‘computers’, who undertook University of Melbourne Council that computing. By then Harry Messel, this work for the (overwhelmingly the hydrogen bomb could not have newly appointed physics professor at male) researchers. Cherry had been been developed without computers, the University of Sydney, was raising sending computations to Sydney to be for it would have been impossible for funds for a more powerful computer, run on the Mark 1 and recognised the humans to undertake the complex to undertake research in nuclear benefits of building a computational calculations: ‘It is simply a device physics. CSIR offered the Mark 1 to capacity in Melbourne, but he also which is absolutely essential today in interested parties. foresaw that digital computing was mathematics, theoretical physics and Thomas Cherry, the University about to transform many fields of statistics’.4 of Melbourne’s professor of research. Cherry forged an alliance It was never going to be mathematics, jumped at the with his Melbourne colleague Leslie straightforward to gain approval opportunity. Up to now, complex Martin, professor of physics, who for such a project in a university mathematical and statistical also saw the benefits of a machine that was operating a deficit budget calculations in the university required that could perform calculations and struggling to invest in research mechanical calculators or logarithm 1,000 times faster than a human equipment. Cherry and Martin

Richard Gillespie Collecting computers—and computer users 13 An early printed circuit board, made by Jurij Semkiw to extend CSIRAC’s disk store, 1961. CSIRAC036, School of Computing and Information Systems Collection, University of Melbourne.

argued that the Mark 1 was worth the operating memory, and construct Preserving CSIRAC £50,000 to £100,000 and available a new magnetic-disk store. After Fortunately, the operating team on permanent loan at no cost. 12 months of testing and technical understood CSIRAC’s historical Additional staff and operational upgrades, the renamed CSIRAC significance. Whereas the other expenses would cost £6,500 a year, (Commonwealth Scientific and first-generation computers in Britain but CSIRO had promised half that Industrial Research Organisation and the United States were scrapped, amount for three years, provided it Automatic Computer) was ready or only parts retained, in 1964 the could continue to use the machine to meet the needs of university complete units of CSIRAC were for its own projects. Martin researchers and external users. donated to the Institute of Applied anticipated, rather optimistically, Over the next eight years, from Science (now part of Museums that the computer could earn £2,000 June 1956 to November 1964, Victoria). As a result, CSIRAC is the from external users, so that the real CSIRAC operated for some 30,000 only intact first-generation computer charge to university funds would hours, with maintenance accounting surviving anywhere in the world.6 The only be £1,000 a year. (University for only 10 per cent of that time. This computer remained in storage until researchers and administrators will was an extraordinary performance CSIRAC’s original designer, Trevor appreciate this financial sleight- for a first-generation computer, and Pearcey, became dean of the School of of-hand.) Council approved the testimony to the skills and dedication Computing and Information Systems funding and the Mark 1 was of the operating team, which over at Caulfield Institute of Technology; carefully dismantled, loaded onto that period included Frank Hirst, in 1980 he arranged for the display a truck, covered with a tarpaulin, maintenance engineer Ron Bowles, of CSIRAC at Caulfield, where it and transported down the Hume Jurij Semkiw, Peter Thorne and Kay remained until 1992, when it returned Highway in June 1955. Thorne. CSIRAC was available for to storage. Once again, a former Physics lecturer Dr Frank Hirst use by university staff and students, CSIRAC operator came to the had been put in charge of the new CSIRO researchers, and government rescue. Peter Thorne, by then head of Computation Department, which departments; in addition, income the Department of Computer Science was overseen by Thomas Cherry and was generated from business users, at the University of Melbourne, Leslie Martin. Hirst later recalled: although rather more modestly than arranged in 1996 for CSIRAC to be ‘We had only one committee Martin had promised! Some 700 displayed in the School of Computer meeting of five minutes duration projects were undertaken, including Science in Bouverie Street, to mark when, by chance, the two Professors physics computations, processing of the 40th anniversary of CSIRAC’s were visiting the Computation survey data, solar radiation tables, operation in Melbourne. Department at the same time’.5 loan repayment tables, life insurance This event triggered a more Hirst had to oversee the re-assembly tables, structural engineering for systematic documentation of of the large cabinets (containing buildings (including ICI House), CSIRAC’s history, and of computer more than 2,000 vacuum tubes), forestry research, electricity supply science at the university. A CSIRAC redesign the mercury delay lines for and climate research. History Team documented CSIRAC,

14 University of Melbourne Collections, issue 22, June 2018 Right: Country school students visiting CSIRAC on Open Day, 1963 (Victorian Pressphoto). 2017.0071.00480, University of Melbourne Photograph Collection 1857–1990, University of Melbourne Archives.

Below: CSIRAC on display in Melbourne Museum, 2001. The panels have been removed to show the valves and circuitry. Photograph courtesy Museums Victoria.

assisted with the display of CSIRAC preserved and accessible. CSIRAC ‘donger’, used by the engineers to tap at the new Melbourne Museum is currently in storage, with options suspect vacuum tubes when CSIRAC (where it remained on view from for its return to public display being was showing an intermittent fault. 2000 to 2017), re-assembled considered. Tapping a faulty tube could make the archives from departmental records As the Computation Laboratory fault permanent, thereby identifying and personal papers, recorded oral door and the rich history of CSIRAC the tube responsible. Sadly, not in histories and video interviews, suggest, computer collections, when the collection is the electric jug used and published a book to celebrate appropriately documented, can for morning tea, which could cause CSIRAC’s achievements.7 John encompass far more than a technical a power spike that erased CSIRAC’s Spencer wrote an emulator program history represented by motherboards memory during a computational run. so that CSIRAC’s programs could and technical specifications. be run on a modern computer, and Sometimes the most peripheral and Computing after CSIRAC the team helped composer Paul incidental items can capture the The collection now maintained Doornbusch recreate CSIRAC’s flavour of working with a computer at by the School of Computing and original music. Through close a particular moment. The CSIRAC Information Systems documents collaboration with Museums Victoria, collection includes a puzzling significant aspects of the history the History Team has ensured that wooden stick with a rubber stopper of computing at the university. In the rich history of CSIRAC is attached: this was the CSIRAC addition to teaching and research, the department provided central computing services for the entire university; a Computer Centre was eventually split off from the teaching department in 1969. As CSIRAC was reaching the end of its life, in 1964 the university invested in an IBM 7044, a second-generation computer with transistors, ferrite-core memory and ten large magnetic-tape drives (those spinning tape drives that have become the iconic representation of 1960s computers, beloved of film directors). This was a very different machine from CSIRAC. The IBM 7044 was a plug-in machine, costing at the time a staggering $1 million, and

Richard Gillespie Collecting computers—and computer users 15 IBM (International Business Machines Corporation, USA, est. 1911), IBM 7044 operating console, 1964. The colourful ‘COMPUTER’ lettering was added by a university staff member, above the formal asset number. Hrd0028, School of Computing and Information Systems Collection, University of Melbourne.

with a service contract that precluded department or research team. In adapted to specific teaching or tinkering. Whereas computer users 1968 the Computation Department research requirements. Jurij Semkiw had driven CSIRAC from the purchased a Digital Equipment gradually expanded the computer control console, the IBM was a Corporation PDP-8, the first and added peripherals as funds batch-processing mainframe; users successful minicomputer (more than became available; these included a brought their punch cards for offline 50,000 were sold worldwide). This state-of-the-art controller to connect spooling, then returned hours later to was a scalable, benchtop computer the computer to a disk drive, and a collect the results of their job. At first that could fit in the lab and be high-speed data channel to connect the computer was housed at IBM’s to the IBM 7044. The PDP-8 is a offices in St Kilda; it then moved to highly significant computer in the the ground floor of the Architecture department’s history, used by staff Building. Engineering student Peter and graduate students for many of Weymouth recalls ‘steaming up the their research projects. window of the Architecture building, Revealingly, the Computation peering in to see if my job had been Department’s PDP-8 was not run … Nine times out of ten the the first of its type to arrive at the printout would say “Error nnn—Job university. Dr David Dewhurst, a Cancelled” and it would be back to reader in biophysics in the Physiology the drawing board’.8 Postgraduate Department, had purchased the students using the IBM would gather first PDP-8 in Australia in 1966. at the key-punch room to enter and Dewhurst was a pioneer in the collect data, and the room became a field of biomedical engineering, forum for the exchange of ideas and and the PDP-8 allowed him to approaches to computing problems. undertake real-time acquisition and The computing collection includes processing of physiological data the operating console of the IBM from his experiments on motor 7044 (possibly the only one surviving activity. Dewhurst became an expert internationally), the memory core programmer, opening up new and the programming and operating research fields for fellow researchers manuals. and graduate students. Dewhurst’s While the mainframe computers lab moved to Electrical Engineering in the Computer Centre continued in 1975, where he and his graduate to be funded and managed centrally, students undertook the electronic researchers were discovering engineering for the development the benefits of purchasing a of the bionic ear, as part of Graeme minicomputer for their own Clark’s team in the Otolaryngology

16 University of Melbourne Collections, issue 22, June 2018 IBM (International Business Machines Corporation, USA, est. 1911), IBM 7044 magnetic-core memory, 1964. This memory- storage technology would be supplanted by semiconductor memory chips in the mid-1970s. Hrd0015, School of Computing and Information Systems Collection, University of Melbourne.

Department. Ian Forster and Jim Patrick designed the electronic circuits for the implantable chip, while Peter Seligman and Jim Patrick worked with Dewhurst on the development of speech- processing software. The PDP-8 is now a particularly significant item in the Electrical and Electronic Engineering Department’s Collection.9 Other minicomputers began to appear around the university: the electron microscopy group in Physics purchased a PDP-8, as did Dr Alan Head at the CSIRO Tribophysics Division at the university. Professor David Caro in Physics purchased a PDP-9 to control the cyclotron; he reputedly thought he had discovered a new element until it became apparent that the orange paint on the computer was slightly radioactive.10 Thus by the end of the 1960s computing was no longer the province of a single department and the university’s central Computer Centre. Decentralisation was not necessarily approved; when an edict was issued that all computers were to be centralised in the Computer Centre, research teams took to describing their new minicomputers as ‘data acquisition systems’. The student demand for access to computers exploded in the 1970s,

Richard Gillespie Collecting computers—and computer users 17 Right: Digital Equipment Corporation (USA, 1950s–90s), PDP-8 minicomputer, 1968. Hrd0025, School of Computing and Information Systems Collection, University of Melbourne.

Below: Digital Equipment Corporation (USA, 1950s–90s), Programs for the PDP-8 minicomputer, 1965–68. CScM0018, School of Computing and Information Systems Collection, University of Melbourne. Punched tape had been used since the 19th century to program weaving looms and send telegraph messages; it was adopted as the standard means to store programs and data for early generations of computers.

ranging from postgraduate research to cards (marked with pencils), or Innovation and undergraduates taking introductory cards with push-out chads (using a commercialisation courses in programming. Graham paperclip). The cards could then be In 1975, Peter Poole commenced Menhennitt recalls ‘standing in read by an optical card-reader and as foundation professor of the line for access to the Interdata 7/16 processed by a locally manufactured renamed Department of Computer computers’.11 In 1974 the department minicomputer. Peter Thorne was able Science, and it grew rapidly in terms adopted MONECS, a low-cost to use such a system to teach a course of staff, students, courses, research system for teaching programming in computing at Pentridge Prison; and equipment. With Poole’s that had been developed by Monash he would enjoy telling his students encouragement, the department also University, in which students back on campus that he had just been undertook significant commercial prepared programs on mark-sense released from prison. ventures. Poole was a strong advocate

18 University of Melbourne Collections, issue 22, June 2018 Apple Inc. (USA, est. 1976), Apple Macintosh personal computer, 1984. Hrd0048, School of Computing and Information Systems Collection, University of Melbourne.

for using open-source operating venture between the university and a 1980s the department used Apple systems, such as Unix, rather than local manufacturer, and although the Macintosh computers in its teaching the proprietary systems of computer project was short lived, the computers labs, but these came with very little manufacturers. Technical staff were used in the department and capacity to be networked to a central designed a computer that would other universities for teaching and computer. Technical staff, including take advantage of Unix software research. Jurij Semkiw, Robert Elz and David and be sufficiently robust and cheap The technical and commercial Hornsby, developed a communications for educational and commercial lessons learnt from this project were bridge based on Unison technology applications. The Unison computer important in shaping subsequent that could link the Macs to Unix was developed as a commercial commercial ventures. In the early servers. A local manufacturer, Webster

Richard Gillespie Collecting computers—and computer users 19 Webster Computer Corporation (Australia, 1970–2016, previously Webster Electronics Pty Ltd), Multigate network terminal, 1988. Hrd0026, School of Computing and Information Systems Collection, University of Melbourne.

Computer Corporation, was licensed The most successful commercial The collection today to manufacture and distribute the venture was the least expected, yet it At the time of writing, the School Multigate network equipment, confirms the department’s international of Computing and Information which was sold widely in Australia perspectives and leadership. Robert Systems Collection comprises and the United States, including Elz had been appointed in the late some 100 objects, including entire to the head office of Apple. The 1970s to facilitate the implementation computers, control panels, processors, university and department received of Unix on the department’s new peripherals such as disk drives and substantial royalties from the licence VAX 11/780 computer, working printers, program tapes and media. agreement. closely with the Unix group at There is also an extensive archive, Other commercial ventures University of California Berkeley. comprising some 900 documents and emerged from software develop- Elz established electronic mail links archival files, including computer ment. Poole had established with Berkeley and the rest of the operating and software manuals, a Development Group in the world during the 1980s, managing a department reports, student research department, which worked with computer called ‘munnari’ to provide reports and photographs. Selected the National Museum of Victoria access for computer science staff items are displayed on Levels 7 and 8 (now Museums Victoria) to develop and other researchers in Australian of the Doug McDonnell Building a database management system for universities; this would evolve into the on the Parkville campus, along museum collections, built around hub for the Australian Academic and with graphic panels celebrating the innovative search algorithms. The Research Network in 1990.12 school’s history and achievements. Titan database was released in 1984, One of Elz’s administrative tasks The collection is catalogued on and Knowledge Engineering Pty was managing internet domain name Microsoft Access software and Ltd was spun off as an independent registrations in Australia, and he stored in the Doug McDonnell company. The database, now called developed ‘reasonableness’ criteria for Building. The CSIRAC collection at EMu (Electronic Museum) and registration that prevented the kind of Museums Victoria comprises some owned by the Swedish Axiell domain name sitting and speculation 300 objects, 300 images and more Group, is used widely by major that occurred overseas. As the internet than 1,000 documents; the collection museums and archives around the expanded, the carefully managed can be explored online at collections. world, including the Smithsonian Australian domain system was publicly museumvictoria.com.au/articles/3145. Institution in Washington DC and floated as Melbourne IT in 1999, The School of Computing and the Natural History Museum in raising almost $80 million for the Information Systems has made London. Appropriately, it is used at university. It is perhaps ironic that this a conscious effort to develop a the university to manage collections hugely successful commercialisation collection that documents its history. at the Ian Potter Museum of Art, is not represented in the collection; all The rapid technological change that University of Melbourne Archives, that has survived is a photograph of a has occurred in computing over the Baillieu Library Print Collection and modem to the munnari server from past 60 years has prompted academic Grainger Museum. the early 1980s. and technical staff to hold on to

20 University of Melbourne Collections, issue 22, June 2018 Elxsi (USA, est. 1979, now Tata Elxsi), Elxsi 6400 mini-supercomputer’s large printed circuit board, densely populated with chips, mid-1980s. Hrd0030, School of Computing and Information Systems Collection, University of Melbourne.

significant items. Most critically, a Yet inevitably there are weaknesses infrastructure and the staff who core of staff recognised the national and notable gaps in the collection maintain and support research and international significance of of artefacts, documents, archives platforms. CSIRAC, and worked tirelessly to and photos. The last three decades There is a larger challenge. preserve and document the computer are not as well documented as The history of computing at the and its history. The CSIRAC History the earlier years. The work of the university has spread out from Team, led by Professor Peter Thorne, Information Systems Department, the Computation Laboratory and has collaborated with university a separate department from its Computer Centre to virtually every colleagues, Museums Victoria, the inception in 1995 until the creation department and centre, and through Australian Computer Society and the of the integrated Department of cross-institutional and government Pearcey Foundation. The CSIRAC Computing and Information Systems collaborations such as the Victorian project naturally flowed into in 2012, is notably absent. Missing Life Sciences Computation Initiative documentation of the wider history of too are collections documenting the (now Melbourne Bioinformatics). computing at the university.13 university’s centralised computer Indeed, just as CSIRAC and its

Richard Gillespie Collecting computers—and computer users 21 associated archives are now part of Dr Richard Gillespie is an honorary 9 Veronica Bondarew and Peter Seligman, The Museums Victoria’s collection, the principal fellow in the School of Historical Cochlear story, Melbourne: CSIRO Publishing, and Philosophical Studies at the University 2012, pp. 32–4; Graeme M. Clark, Sounds from university’s computing ‘collection’ is of Melbourne, and formerly head of the silence: Graeme Clark and the bionic ear story, more accurately characterised as a Humanities Department at Museums Victoria. Sydney: Allen & Unwin, 2000, pp. 76–88; distributed collection, spread across He has written widely on the history of science Richard L.G. Kirsner and John S. McKenzie, and technology, including The Great Melbourne ‘David Dewhurst: Biomedical engineer and the university and encompassing Telescope (Melbourne, 2013). IFMBE pioneer’, in Laura M. Roa (ed.), CSIRAC. XIII Mediterranean Conference on Medical and Acknowledgements: The author acknowledges Biological Engineering and Computing 2013, Recording and preserving the the advice and support of Robin Evans, Tony IFMBE Proceedings, vol. 41, 2014, pp. 9–12. significant aspects of computing at Klein, Bernard Meade, Alistair Moffat, Michael 10 Max Burnet, ‘CSIRAC, Melbourne University Poloni, Jurij Semkiw, Graeme Shanks, Liz the university now require a more and Digital: A long relationship’, in McCann Sonenberg, Peter Thorne and Justin Zobel. strategic approach, one that would and Thorne, The last of the first, p. 98. identify objects, images and stories 11 Graham Menhennitt, quoted in ‘History of computing in the department’, www.cis. across all research fields and research unimelb.edu.au/about/history/computing/, platforms. We now need to think 1 Doug McCann and Peter Thorne, The viewed 28 February 2018. of the collection as a distributed last of the first: CSIRAC: Australia’s first 12 Glenda Korporaal, AARNET: 20 years of the collection that encompasses the computer, Department of Computer Science internet in Australia, 1989–2009, Sydney: and Software Engineering, University of AARNET, 2009. perspectives and endeavours of many Melbourne, 2000, p. 46. 13 Suelette Dreyfus, ‘The first years 1996–2004’ different parts of the university. 2 ‘He’s a real Mr. Felix!’, Argus, 4 May 1956, p. 3. [History of the Department of Information Computing can be a case study 3 Paul Doornbusch, The music of CSIRAC: Systems], School of Computing and Australia’s first computer music, Melbourne: Information Systems, Melbourne School of for analysing the transformation Common Ground Publishing, 2005, pp. 23–7. Engineering, University of Melbourne, 2004, of teaching and research at the 4 ‘Giant “brain” sets ’varsity a problem’, Age, www.cis.unimelb.edu.au/about/history/pdf/ university, encompassing discovery 7 December 1954, p. 3. department-history.pdf, viewed 28 February and innovation by individuals and 5 Quoted in McCann and Thorne, The last of 2018; Alistair Moffat, ‘Fifty years of the first, p. 39. computing at the University of Melbourne’, teams, but also the resources required, 6 See Heritage Council Victoria, ‘CSIRAC Melbourne School of Engineering, University the technical support, the institutional (Commonwealth Scientific and Industrial of Melbourne, 2006, people.eng.unimelb.edu. climate and strategic commitment. Research Organisation Automatic Computer): au/ammoffat/fifty-years/mof06history.pdf, Statement of significance’, vhd.heritagecouncil. viewed 28 February 2018; ‘Computing at the This is a story of computer users vic.gov.au/places/114928, viewed 16 April University of Melbourne, 1956–2015’, School rather than computers. 2018. of Computing and Information Systems, 7 McCann and Thorne, The last of the first. Melbourne School of Engineering, University See also Maston Beard and Trevor Pearcey, of Melbourne, www.cis.unimelb.edu.au/about/ ‘The genesis of an early stored-program history/, viewed 28 February 2018. University For further information and access computer: CSIRAC’, Annals of the History of Melbourne Archives also holds records to the School of Computing and of Computing, vol. 6, no. 2, 1984, pp. 106–15. and artefacts relating to the early history of Information Systems Collection, 8 Peter Weymouth, ‘After CSIRAC: A non- computing; see Mary Hurley McGillivray, please contact Dr Richard Gillespie user’s view, 1970–2000’, Museums Victoria, ‘Early computing’, University of Melbourne on [email protected]. 2007, collections.museumvictoria.com.au/ Archives, archives.unimelb.edu.au/resources/ articles/1349, viewed 28 February 2018. subject_guides/early-computing.

22 University of Melbourne Collections, issue 22, June 2018