• Expert SAP HR and / or Payroll...
• GIS / SPATIAL DEVELOPER - CIRCA...
• PROJECT QUALITY ANALYST
• Senior Architect - 12 mth contract
• GROUP MANAGER INFORMATION SYSTEMS

John Bennett honoured with 2004 Pearcey Medal.

Peter Davidson, Information Age

14/12/2004 11:43:21

Presented by Sen Helen Coonan, Minister for Communication, Information Technology and the Arts in Sydney on November 4, the medal commemorates the work of Trevor Pearcey, a contemporary of Bennett's and builder of CSIRAC, Australia's first computer.

With a degree in civil engineering and four years' experience in radar with the RAAF during the war, his service record and age (he was unde 21 when he graduated) gave him access to any university course. He chose electrical engineering, mathematics and physics.

A summer vacation stint at CSIR (now CSIRO) in its electro-technology division brought him to first contact with Pearcey, and shaped the rest of his career. Later the pair would be instrumental in establishing the ACS's state branches, and later the national organisation of which Bennett was foundation president.

An application to go to England to join the team working at the National Physical Laboratory on ACE (Automatic Computing Engine) fortuitously found its way to Dr Douglas Hartree at the Cambridge Mathematical Laboratory. He passed it to his director, Maurice Wilkes, who was just embarking on his EDSAC project to build the world's first stored program computer there. (See Sir Maurice Wilkes, going strong at 91 in this issue.)

Joining the team as its first graduate student in 1947, he made a significant contribution to the project before being recruited in 1950 by Ferranti Ltd to work on its Ferranti Mark 1 commercial computer.

He returned to Australian in 1955 to become the first professor in computing science at the Basser Laboratory at Sydney University where he led the team that built SILLIAC to a design based on the University of Illinois' ILLIAC machine.

He also worked to develop Australia's ICT knowledge base by fostering relationships with industry and generating a growing flow of new graduates. By his retirement in 1986 aged 65, 2000 students had graduated from Basser, and another 4000 from all disciplines had completed courses there.

His work since with the International Federation for Information Processing (IFIP), continuing involvement with Sydney University, particularly in post-graduate affairs, and close involvement with the Australian Computer Society of which he is a Fellow has kept him busy.

The following is his speech in reply to the presentation:

Early Australian network development "Trevor and I had parallel careers. After graduating early in the war, Trevor spent a few years at the Radar Research Establishment Great Malvern UK and I spent my war years on the operational side of RAAF Radar which included an excellent radiophysics course at the University of Sydney Physics School. On my discharge at the end of 1945 I underwent further electrical engineering training under the Commonwealth Reconstruction and Training Scheme and for experience spent time with CSIR (now CSIRO). Trevor had just joined this group from the UK, and went on to design and build a machine which became CSIRAC.

At the time I was planning future electrical power networks in the Brisbane area involving lengthy repetitive calculations, and heard a talk on the UK ACE computer then being planned. My efforts to join a UK project of this type resulted in my joining the Cambridge EDSAC group. EDSAC was the first electronic computer in regular operation (in May 1949) and I was responsible for designing, constructing and testing its Main Control Unit which extracted instructions from the store and carried them out. I also designed, constructed and tested the bootstrap facility which started up the machine.

Trevor's and my work at this early stage is summarised in "Computing in Australia", a 1994 ACS publication. After Cambridge, I built up the Ferranti group for the Ferranti Mk1, redesigned the order code for the Mk1, and built up a design group in London which resulted in PERSEUS, a machine for the insurance industry. The first Perseus was installed in South Africa at the time I returned to Australia to join the SILLIAC group - SILLIAC was then under construction, mainly by STC to our ILLIAC-based design.

It was succeeded by the English Electric KDF9 which was interconnected with a CDC machine and SILLIAC to form the first Australian computer network, with over a dozen keyboards, which became operational in 1967.

In the early 60s, Pearcey's group was developing a computer network (CSIRONET) using magnetic tape as a basis for interchanging programs and data. Two levels of design included a lower level like the University of Sydney CDC machine.

The world scene and the future

Networking has now become worldwide practice, culminating in Internet, engineered by Sir Tim Berners-Lee (his mother was a star programmer in my Ferranti team; he was knighted in July this year). By the end of 2000, the Internet had over 25 million Web sites. Details will be found in the October issue of MIT Technology Review and comment on its growth in September/October issue of IEEE Internet Computing.

Perhaps the best overall summary of the computer networking position will be found in October/November issue of The Economist. Now three-quarters of spending is on adjusting existing systems - the proportional spending on new equipment is shrinking. Inevitably, as with the development of the motor car, more and more is being handled automatically with the creation of more layers of abstraction. Inevitably many languages will become extinct. With all the other languages, remember the one spirit left in Pandora's box was Hope.

Perhaps the image of LEGO blocks is relevant - each block being a complicated software unit. And the blocks will become larger - but inter-connectable. LEGO blocks can be complicated inside but the interconnections simple. And wires will not be necessary - wireless technology (WiMag) will be available next year, with 50km range.

For multiple developers, standards are necessary for interconnecting if the digital home is to arrive. But complete Artificial Intelligence (AI) will continue to elude us. VoIP (Voice over Internet Protocol) telephony) is a compromise. The OK word is now SIP (Session Initiation Protocol) to integrate voice and other software.

The overall picture of progress is to consider non-consumers in the near future. Look for simplicity and affordability, so that more time and energy is available to process the infinite complexities of business and life in general."