Tag Archives: Mikhail Gorbachev

Soviet computer science: once ahead of the west?

Coat of arms of the Union of Soviet Socialist ...

In my old blog I made a point that I still believe – that the key factor in the downfall of the Soviet Union was not pressure from the arms race but the complete failure of the system: once the country was led by leaders, like Mikhail Gorbachev and his team, who were open to learning from the west, then the Soviet system was doomed, because why would any enlightened leader what to keep a system that was obviously a fiasco? (I recommend Archie Brown‘s The Rise and Fall of Communism if you want to know more.)

But the collapse of the Soviet Union was just that – a collapse. There was no “transition” as there was in the case of central Europe - this obituary of Vaclav Havel in The Economist is a moving and powerful reminder of those times – the place fell apart, life expectancies and incomes crashed and scientists fled to anywhere they could make a living – something that continues to worry those who are concerned about nuclear and germ warfare proliferation.

Science in the USSR was dominated by the demands of weapon production and even the most brilliant scientists and mathematicians were not exempted from repression, as Andrei Sakharov‘s case showed  – but there was also a very strong culture of maths for maths’s sake – as this article in The Wall Street Journal relates.

From an early age those who showed promise in maths were hot housed in special schools (The Honors Class: Hilbert’s Problems and Their Solvers relates how this, authoritarian system shaped the life of Yuri Matiyasevich, one of the co-solvers of Hilbert’s tenth problem) – and those who showed promise were both privileged and tightly controlled.

Why am I writing all this now? Because I have just obtained a copy of the out of print Stochastic Analysis of Computer Storage (I think I got the last cheap-ish copy) by the Soviet computer scientists O. I. Aven and Y. A. Kogan, and AT&T’s E. G. Coffman and was rather shocked to discover it contained what looks to me like an outline of the LRU-k page replacement algorithm a full six years ahead of O’Neil, O’Neil and Welkum’s 1993 paper “The Iru-k page replacement algorithm for database disk buffering.” (NB I am not for an instant suggesting that the 1993 paper was cribbed from the earlier work)

Even more remarkably, the book refers to papers written in Russian and published in Minsk in 1977 as its source.