• Project Manager - Trading...
• Support Analyst Programmer
• Business Analyst - Integration
• SAS Analyst
• MANAGER STRATEGIC INDUSTRY...

John Bennett: Educating the technology generation

Peter Davidson, Information Age

12/08/2003 15:55:12

Emperor Yu was supposedly strolling along the Yellow River one day around 2200 BC when he spotted a turtle: its shell had a series of dots within squares. To Yu's amazement, each row of squares contained 15 dots, as did the columns and diagonals. When he added any two cells opposite along a line through the centre square, like 2 and 8, he always arrived at 10. The turtle, unwitting inspirer of the ''Yu'' square, went on to a life of courtly comfort and fame.

For some in their 80s, a few crosswords might offer enough cerebral gymnastics to sustain their mental fitness.

Not for ACS founder and Emeritus Professor of Computer Science John Bennett however: he goes to Clifford Pickover's "The Zen of Magic Squares, Circles and Stars" to get into an intellectual wrestling match with its mathematical permutations of magic squares, recreational maths being an elixir to this 82-year-old pioneer.

Sitting in the sharp winter sun filling his lounge room which looks across Sydney's North Harbour to the horizon, conversation about his co-founding of the NSW Computer Society and efforts to create a national society, flows easily.

Technical periodicals and books are stacked at hand awaiting his attention, around them scraps of paper carrying his magic square jottings, and talk soon turns to grid computing as he recalls a recent PhD seminar he attended on the topic at the University of Sydney to listen and offer "a bit of advice".

He tries to attend as many as he can at the university where he joined Harry Messel's School of Physics group in 1956 to head the operational side of SILLIAC (built by Brian Swire and Barry de Ferranti), and later became Australia's first Professor of Computer Science, heading up the Basser Department there.

With a degree in civil engineering and four years' experience in ground radar with the RAAF at the end of WW2, he was entitled, as he was under 21 when he graduated and by dint of his war service, to enrol in any university course. He opted for further studies in electrical engineering, physics and mathematics.

A summer stint at the electrotechnology division of the Council for Scientific and Industrial Research in 1945/6 (now CSIRO) where David Myers was working on analogue and digital computing brought him in touch with the scientific discipline on which half a century of his pioneering academic life would be founded.

It was there that he met Trevor Pearcey who went on to design and build CSIRAC in 1951, Australia's first stored program computer.

Their meeting also spawned a drive by the pair to found computer societies initially in NSW, Victoria and Canberra, and in 1966, the Australian Computer Society with Bennett as its first president and Pearcey as vice president (and its president in 1967).

"To me, it was a matter of spreading the gospel; I knew that computers were here to stay and I knew that they offered enormous possibilities for taking the grind out of repetitive calculations," Bennett said.

He knew about repetitive calculation: joining the Brisbane City Electric Light Company in 1946, his work on its power distribution network brought plenty of it but a chance hearing of a radio talk about the ACE computer being developed by the National Physical Laboratory outside London offered a solution to his mathematical grind, and would set his career path.

He applied to the nearby London University Imperial College's electrical engineering department for a research studentship. Fortuitously his application was passed to Douglas Hartree of the Cambridge University Mathematical Laboratory, who arranged for him to join the Electronic Delay Storage Automatic Calculator (EDSAC) development team, headed by lab director Maurice Wilkes (now 90 and knighted), in 1947.

War surplus technology

EDSAC would become the first stored program electronic computer in regular operational use "We used war surplus valves bought by the pound" and I/O equipment based on uniselectors scrounged from the Cambridge Post Office to build EDSAC in what had previously been the dissecting room of Cambridge's anatomy school.

"It had a useful goods lift big enough for two cadavers, but on the other hand the smell of formalin-soaked floorboards pervaded everything in hot weather."

It was on EDSAC, and later in Manchester, that Bennett carried out what were to be the first structural engineering calculations on an electronic computer, resolving problems like flutter in aircraft airfoils, the work forming the basis of his PhD.

The importance of this work was recognised by many who followed. In his history of computing and structural engineering, Dr R K Livesy wrote: "what made it significant was Bennett's realisation of one of the most important lessons of computing - that one must be willing to adopt new methods and not merely to take over existing techniques of manual calculation."

Later joining Ferranti Ltd in Manchester, the young and free-ranging computer industry offered a variety of pursuits for Bennett including various marketing assignments, machine specification and running a programming group initially including six women.

Moving to Ferranti's London Computer Laboratories in 1953, he found himself working within 100 metres of the site of Charles Babbage's house. Known as the "father of computing", the 19th century mathematician's Difference Engine Number 1 was the world's first successful automatic calculator.

(Babbage later developed his Analytical Engine [1856]), which was intended as a general symbol manipulator, and had some of the characteristics of today's computers. He died in 1871, embittered and disappointed by government failure to support the production of his machines…)Bennett worked in a team led by Bill Elliott, and which included Charles Owen whose plug-in units were designed so that logical design of complete computers using them could be done by non-engineers. Owen went on to design the IBM 360/30.

"Whatever we touched was new; it gives you a great lift. We weren't fully aware of what we were pioneering. We knew we had the best way but we weren't doing it to convert people - we were doing it because it was a new tool which should get used.

"We knew we were ploughing new ground. We were top of the heap."

Back in Australia he joined Harry Messel's School of Physics group in 1956 to head operations on SILLIAC, the Sydney version of ILLIAC, the University of Illinois Automatic Computer. Considerably faster than any machine then available commercially, its individual sections engineered by Brian Swire were built under contract by STC, making it in a sense the first computer built by Australian industry.

As well as taking charge of the software needed to run SILLIAC, he began to develop associated courses. Not part of any degree sequence, they suffered a fall off in numbers when examinations in credit subjects were scheduled.

In 1959 a formal postgraduate Diploma in Numerical Analysis and Automatic Computing was created, later changed to Diploma in Computer Science.

Still not readily accepted, computing's development received a fillip from a grant by BHP in the 1960s which allowed the recruitment of specialised lecturers with experience in the practical aspects of commercial information processing, and in 1966 an Honours year was introduced.

Melbourne Cup funding

At the same time SILLIAC was decommissioned but not without nostalgic affection: "SILLIAC had two heat exchangers, one at each end of the machine. Each had just sufficient capacity to cool six bottles of beer."

External grants were important to the fledgling Basser Computing Department, named after Adolph Basser, one of the university's benefactors who supported SILLIAC's development from the 50,000 pounds he'd won on the Melbourne Cup in the 50s.

After its split in 1972 into teaching/research and service computing, Professor Bennett headed the Basser Department of Computer Science and Robert Donnelly the Basser Computing Centre.

The historical circle continued to turn in 1979 with the department's move to its present home in the Madsen Building which formerly housed the CSIRO's National Measurement Laboratory where CSIRAC had been built.

Fostering relationships with industry and striving to ensure a flow of graduates skilled to meet its demands has been a cornerstone of his tenure, reinforced by his involvement in establishing the Research Foundation for Information Technology at the university in 1981.

"I'm convinced that what we did in the early days in training a lot of people and letting them loose on the community did much more good for computing than all our technical expertise."

In 1986 the Vice-Chancellor of the day halved the student intake and aged 65, John Bennett retired, leaving a huge legacy in Australia's knowledge pool. At that time, about 2000 students had graduated from Basser, and another 4000 from all disciplines had completed courses there.

The Australian total of computer science majors up to the end of 1986 was about 4000 from all universities and 2000 from Colleges of Advanced Education.

Six holders of computer science professorships at the time in Australia and New Zealand had come from there, and its enrolment of Honours and PhD students was the highest in Australia. The flow has continued.

Back to the present

And so the conversation turns back to the present and the way in which grid computing will carry our technology forward "from the crest of a wave". I ask if I can take his photo to accompany this article.

"If you wish", but not before he starts a lengthy cross examination on the technical aspects of my little Nikon digital, scribbling details of resolution, weight, optical quality and price in a notebook pulled from his pocket. "I'll need one for the grandchildren."

You can take the man out of technology, but…Extracts from the University of Sydney News are gratefully acknowledged.

[breakout box] In the first of an occasional series of articles on the work of ACS members in the pioneering of Australian ICT, IA discusses the past and future with ACS founder John Bennett, AO.