Ken Robinson - A Remembrance

Many, many people would have led different lives had it not been for Ken Robinson: so many dimensions in the one person, so much to miss — and so much for us all to remember and to celebrate.

| 16 Feb 2021
Ken Robinson

This article is sourced from the British Computer Society, FACS FACTS Newsletter, Issue 2021- 1 February 2021, pages 26 to 31.

Kenneth Arthur (Ken) Robinson joined the staff at the University of New South Wales in the mid1960s. Always a high achiever, Ken had been dux of primary school, dux of secondary school and had topped his joint Bachelor of Science/Bachelor of Engineering degree course at the University of Sydney. 

When Ken moved to UNSW, he at first tutored Mathematics, but by 1968 he had become one of the three systems programmers in charge of the University’s sole general-purpose computer — for which academics used keypunches the size of ironing boards to prepare their IBMFORTRAN 80-column-card decks. They got one run a day, and their data were read from 7- megabyte multi-platter disks — like wedding cakes — carefully lowered and screwed tight into drives that were as big as washing machines. Undergraduates however were not allowed to use the keypunches. Instead they had to buy specially perforated 40-column cards, learn the lettercodes by heart and then poke-out the corresponding holes with a specially bent paperclip. For them, it was only one run a week — and usually resulted in a compile-time syntax error.

So that was where Ken’s Computer-Science career started, at UNSW. He remained there until 2012 and passed away in late 2020. Ken was an engineer at heart, and so had the advantage of an extra perspective of rigour and precision. Over the many decades of his association with UNSW, he was variously lecturer, Head of the Department of Computer Science (later School of Computer Science and Engineering, CSE) and, as Associate Professor, the designer and force behind CSE’s Software-Engineering Programme, the only such programme in Australia with a solid formal-methods core and furthermore one of the foremost of its kind in the world. Ken was the only engineer among his colleagues, and yet, in spite of being surrounded by computer hardware and software, he still loved his pen and pencil — he saw them as craftsmen’s tools.

Through all those years, Ken was the intellectual mentor of hundreds of students, many of whom considered him — at the time and also now — to have been the most inspiring teacher they had ever experienced. He had a habit of not quite giving a direct answer, so that the student had to invest some original thought as well. His lectures were not “performances” (except perhaps occasionally when he had to brush back his long, blond hair as he spoke) but were marked rather by his quiet incisiveness of thought and exposition, placing him among the first four academics at UNSW to receive the “Vice-Chancellor’s Award for Teaching Excellence”. Being in his class transmitted an infectious compulsion to absorb as much from him as one absolutely, possibly could — a kind of explosive exhilaration. One felt “This is how teaching should be.”

Why was Ken so conspicuously excellent? In spite of Australia's distance from the rest of the computer-science world in the early 1970s: journals and conference proceedings arriving through the ordinary post, after months at sea; telephone calls even to neighbouring capital cities in Australia still needing to go via the operator; recent papers from the northern hemisphere requested by aerogramme with the paper arriving (perhaps) in a manila folder weeks later, sometimes addressed to the University of North South West. Nevertheless Ken Robinson put us — himself and his students — right at the very front of the astonishing new developments in formal methods and rigorous computing generally. 

In 1971, Ken’s front-line courses in computer science (in second year because then, at UNSW, Computer Science had no first-year courses), there was ALGOL-W (from Stanford), WATFOR (student FORTRAN from Waterloo), Plago (student PL/1 from Brooklyn), SNOBOL (originally from Bell Labs), and even IBM/360 assembly language — for the last, using an assembler that Ken wrote himself. (The IBM version was too slow and unforgiving for student use.) All that in one year.

From “load register from offset(base)” all the way up to Floyd-Hoare-Dijkstra high-level reasoning about programs at the source level: their abstractions, refinements, specifications, implementations. All of that was brought to UNSW, then promoted and further developed by Ken — who understood, and propagated the truth that to be effective and professional you really had to understand things from top to bottom, and back again. You can’t abstract properly unless you know what you’re abstracting from and (therefore) understand the cost of not doing so. And you can’t refine unless you know the environment you are aiming for. Ken worked and taught at the crossroads of elegance and practicality — the things that caught his eye had to be both beautiful and useful.

What about the very latest systematic compiler-writing methods? For those, there was Ken’s fabulously popular third-year course, the first introduction of a compiler subject in Australia. And the astonishing PASCAL language, that mysteriously could compile its own compiler? Obtained directly from ETH Zürich by Ken who again was the first to teach it in Australia. And UNIX, a system recognised by Ken as being “too good to be true”? He wrote to Dennis Ritchie at Bell Labs to acquire a copy, which arrived in 1975. It made UNSW the first University outside the USA to run UNIX as a production facility. And functional programming? There too Ken was the first to use Miranda in Australia as a teaching language. And for the hard-core formalists, a course on Denotational Semantics? The text (Gordon) was “not available” directly from Sydney booksellers and “would take months” to obtain. Ken braved the international telephone operator and ordered 20 of them to be sent directly to Australia from Blackwell’s in Oxford. By airmail.

Ken has often been described by his colleagues as “The Father of Formal Methods in Australia”. The above shows why. However it was not all just one way: Ken’s passion for sharing his expertise in teaching Formal Methods and Software Engineering extended in the other direction too, that is, from the South back to the North. He established a strong connection with Oxford University’s Programming Research Group, and he helped many, many colleagues in the UK and France who were involved in the B community founded by J-R Abrial — now many of them professors themselves. He was limitlessly generous in sharing his resources and giving of his time to talk-through examples that would form good exercises, creating material that showed how he wanted students to learn in depth but also to be able to apply what they learned. He was keen that students should see B and Event-B systems as practical methods with effective tool support, as any engineering discipline must be. And so he incorporated the use of those systems into his SoftwareEngineering stream back in Australia, and made sure that code could be generated from the specifications with ease. He showed the students where formal methods could and should be used in a software-engineering lifecycle through group projects that enabled the students to see and experience their relevance to their own future careers. When the South and North were together at Formal-Methods conferences, Ken was always keen to hear how teaching was progressing “up there” and to learn from the student feedback.

And Ken always encouraged younger academics to care about their students and to take the time to find their own research strengths. He was always available to provide quiet mentoring, and his interest in their careers was long lasting.

As an engineer, Ken was especially passionate about computer science as a branch of engineering and crucially, for him, that included not only the social and methodological aspects of engineering, but also the “mathematics” (think differential equations for Electrical Engineering) that, for computing, was closer to logic. He was not a fan of what he called “high-level hacking” without a solid basis. As the methods and tools for reasoning about programs, and the systems made from them, continued to develop throughout the later ’70s, then ’80s and ’90s, Ken contributed directly to the development and teaching of what is currently the “Rodin” toolkit for constructing, organising and provingcorrect large-scale computer systems. Rodin’s latest release has been dedicated to Ken. His “Software-Engineering Stream” within the UNSW computer-science curriculum started with a Requirements-Analysis project in First Year (Computer Science had a first year by then), followed by a formalisation of its specification in “Event B” (within Rodin) in Second Year, and then a rigorously controlled development trajectory from there via program refinement to actual running code, in Third Year — all of that built on the foundational work of J-R Abrial and others from the early 1980s. 

Indeed, Ken was one of Rodin’s very early adopters, and was not shy in sharing his views, both positive and negative, with the development team. The comments were always constructive, and they stemmed not only from his deep knowledge of model-based formal methods and refinement but also from his focus on effectiveness and usability of formal-methods tools. Many of Rodin’s features resulted from email exchanges between Ken and the tool’s main architects. His use of Rodin in teaching began around 2007. At the first Rodin User and Developer Workshop in Southampton in 2009, he gave an inspiring talk on how he had managed to incorporate formal modelling and proof, supported by Rodin, into the undergraduate software-engineering courses at UNSW. The key message from that talk was the importance of good design and how abstraction and refinement support it. It was very encouraging to see an affirmation of the software-engineering value of tools like Rodin by someone as respected as Ken.

On the more personal side, Ken had an insatiable curiosity about the world, an endless desire to find out more about everything — whether it was how things worked, how people thought or just ideas, history, films, music... and more. Ideas were most important: juggling viewpoints, and finding different ways of approaching problems. His life at university was of course a natural extension of that. Education was one of Ken’s strongest passions and it was through teaching that he transferred his enthusiasm for whatever it was. It was important to him to encourage this curiosity, this thirst for knowledge, in others.

Although Ken dedicated a huge part of his life to the university, to computer science and software engineering, he also had a deep concern for humanity and for the state of the world generally. He was a person of enormous generosity, with whom it was immediately possible to fall into a simple friendship that seemed somehow to have existed already. People were very important to Ken. Those lucky enough to know Ken personally often mention the perpetual twinkle in his eye, the feeling when speaking with him that a smile was always about to break out. He cared deeply about others who were less privileged. He was a determined advocate for the rights of everyone. He was a person of strong principles, and so it was important to him to remain true to those things he valued. It didn’t matter too much what others thought — it was more important to stand up for what you yourself believed in.

Ken was down to earth with absolutely no pretensions, and indeed had an intolerance for those valuing status in whatever form. In the early days, much to the horror of many, he refused to stand for the Australian national anthem at the time, “God save the Queen”. (In Australia it used to be played at the end of concerts, and in cinemas before the beginning of every film!) The frills and glitz of life were irrelevant to him. His working life at university reflected these values and his students were his equals: “Just call me ‘Ken’!”

For Ken everything had to be done, in his words, “properly”. He was a perfectionist, and there was always more work to be done on improving whatever tasks were at hand. And yet at the same time Ken was unbelievably untidy, being described by colleagues (with some understatement) as “a little messy”. But that was just part of his concentrating on what the important things really were. His dishevelled room at UNSW was well known, and in one story he related how the police arrived to investigate a robbery in the building. Apparently they were sure his room had been ransacked. It hadn’t: it was just in its “normal” state, filled with the clutter of ever-increasing papers and student assignments — just as it was at his home, where once the babysitter couldn’t find the telephone buried beneath mounds of books and papers! 

Further outside of academic life, Ken was an explorer of almost everything, with an enthusiasm that carried along with him anyone within range — very often his whole family of five: from bushwalking, to cycling, to woodwork, to photography, to all five living (temporarily and including a 4-week old baby) in a van during a sabbatical in the UK until better accommodation could be found. He loved the outdoors, from where many a funny, and sometimes not-so-funny family story would emerge. There was rarely a bushwalk where the family didn’t get lost taking one of Ken’s so-called “short cuts”. But this was — of course — just part of his finding new and different ways of doing things!

In his early days Ken spent many a weekend fixing and polishing his MG-TD as well as playing the clarinet. He was very musical, and his appreciation extended from early music right through to the contemporary. Bach was a particular favourite, anything from the Amsterdam Loeki Stardust Recorder Quartet’s interpretation of The Art of Fugue (“As precise as a pipe organ!”) to, in more recent years, humming along with Glen Gould playing the Goldberg Variations (which the family included in his funeral service). He also on occasion sang in the chorus for the annual Handel’s Messiah at the Sydney Opera House. Shakespeare was another favourite, whether it was reading his plays or attending performances. Among Ken’s other passions were cooking and breadmaking. Apart from his much-loved Thai and Indian curries, his extraordinary home-made bread –a “perfected” recipe– won many prizes at the annual Sydney Easter Show.

Many, many people would have led different lives had it not been for Ken Robinson: so many dimensions in the one person, so much to miss — and so much for us all to remember and to celebrate.

This article's contributors included:

Jonathan Bowen, Michael Butler, Paul Compton, Ian Hayes, Martin de Groot, Peter Ho, Son Thai Hoang, Carroll Morgan, Maurice Pagnucco, Aaron Quigley, Rosalie Robinson, Claude Sammut, Steve Schneider, Martin Schwenke, Arcot Somya, Bill Stoddart, Jeff Tobias and Helen Treharne.