When I met David Barron, he was wearing pink socks. His explanation was typical of his sense of humour, and of our subsequent research, which centered on taking real measurements. We investigated whether drinking Cinzano made them pinker. By the time we had enough data, though, we could not take the definitive decision, but we went home happier.
That was in January 1973, when David was part of a team of professors from the UK that IBM brought to Cape Town to “Teach the Teachers” computer science. David presented his work on comparative programming languages, explaining so simply and clearly about l-values and r-values, that by the time he got to call-by-name, he still had the large group with him. We had used David’s Assemblers and Loaders and Operating Systems books in my undergraduate classes, so he was already famous, but his work on languages was a revelation.
In 1974, David helped acquire one of the brand new ICL 2970 computers for the University of Southampton, and I came over to work as a PhD student to take those measurements we had discussed:
“investigating actual code produced by the compiler and included in its run-time package; devising ways of determining which inefficiencies are an inherent legacy of the language design, and which are accidents of implementation; and investigating (again quantitatively) the effect of different machine architectures to see if “stack machines” are as well-suited to high-level languages as is commonly maintained.” [Letter to Miss Mullins, 4 June 1974]
David called this kind of work “Software Engineering”, and in true engineering style, we started by building something – a Pascal compiler for the 2970. Our first attempt was to retarget the existing first compiler for the CDC 6600, which had already been done for an ICL 1907. This required getting 9” inch magnetic tapes posted from Zurich, and a trip to Belfast. Then because David and I discovered that we were kindred spirits in loving parties, we organized the very first Pascal symposium (Latin for drinking party) was held at Southampton in March 1977 in the Maths Block and featured such notables as Per Brinch Hansen, Brian Wichmann, and the originator of the Pascal Newsletter, Andy Mickel. For several years, the Newsletter was printed and shipped from Southampton.
David was never scared of taking a new road. In 1975, he decided to introduce Pascal as the language for second year teaching – even though we only had our somewhat rickety compiler. With a course that he put together over the vacation, and with interesting examples for the students to try, the course was an academic success, and led to a book that we wrote jointly on “Advanced Programming”. Morley Sage, then the Director of the Computer Centre, was not so happy, as the compiler used many more machine cycles than WATFOR, but David was determined that students would not continue to be subjected to what Dijkstra has labeled as an “infantile disorder”.
I admired David immensely – for his vision, his clear thinking, his humanity, and his prodigious output of books. I am very proud that he was my PhD supervisor – I was his second graduate – and that when I returned to Southampton as a Reader in 1988, I could call him a friend. I offer my sincere condolences to Nik and Jacky who have lost a dear father. I know that when we all wear bright ties on Friday, David will smile, and know that he made his mark in his own way.
- Dr Judith Bishop, Director Computer Science, Microsoft Research