Peter Landin: a computer scientist who inspired a generation, 5th June 1930 - 3rd June 2009
Richard Bornat · Formal Aspects of Computing · 2009
Peter Landin, who has died recently of prostate cancer, was an outstanding academic computer scientist who set the direction of the subject for his generation.In an astonishing burst of creativity in the 1950s and 1960s he showed that programs could be defined in terms of mathematical functions, how to translate them into functional expressions in the lambda calculus, the creation of Alonso Church in 1936, and how to calculate their meaning with an abstract mathematical machine, prefiguring many of the advances of the next couple of decades.Although later he turned away from computer science towards gay politics, he remained an original and extreme thinker, a complex and intriguing character.An only child born in Sheffield, son of an accountant father disabled in WW1, he was educated at King Edward's Grammar School.During National Service he amused himself by trying to drill a squad to imitate addition in binary arithmetic.At Clare College Cambridge he completed the mathematics degree in two years, then attempted the very difficult part 3, but came away with only a 3rd class degree.The emerging computing industry was a haven for mavericks, though, and he took a job at English Electric, where he cooperated in the development of EE's Deuce from the National Physical Laboratory's early Pilot Ace.His ideas had already turned to the problems of relating the mathematics of logic to the ad-hoc languages invented by programmers.He was involved in the discussions which led to the development of Algol 60, a seminal programming language of the time, though he failed to convince the committee that it should be defined in terms of substitution as in the lambda calculus.His interests drew him into Mervyn Pragnell's discussion group.Mervyn used to prowl bookshops spotting people who were buying books on mathematical logic and invite them to his group which read and discussed the books in a kind of perpetual cycle.Despite the depth of his understanding compared to the rest of the group, Peter was always kind and helpful to novices and stragglers, and even programmers could join and learn.In 1960 he married Hanne and he was taken on as a research assistant by Christopher Strachey, then an independent consultant and later Professor at Oxford, to write a compiler to translate 'autocode', an early programming language, into the machine language of Ferranti's new Orion machine.He envisaged the compiler as an automatic product of the semantics of the autocode, matching its forms to semantic representations of the instructions of the machine, and generating LISP expressions that could be executed, but his attempt was never finished, and it was a long time before such a radical approach became feasible.He was held in awe by the elite