To celebrate the IMA’s 60th anniversary in 2024, a number of articles have been chosen by IMA members and MT readers, who introduce their favourite MT articles and explain why they are important to them. Here Peter Rowlett CMath FIMA, Sheffield Hallam University, introduces …
How to get Meaningless Answers in Scientific Computation (and what to do about it) by Leslie Fox FIMA
Leslie Fox (1918–1992) was influential in the development of numerical analysis as well as involved with the early history of the IMA, its Journal, its Journal of Numerical Analysis and the Numerical Algorithms Group (NAG). The IMA’s Leslie Fox Prize for Numerical Analysis, established in 1985, was named for him.
In this article, Fox argues for the need of expert judgement in working with computers, pointing out that serious errors can and do occur in practice – he claims ‘about 80 per cent. of all the results printed from the computer are in error’. He aims to illustrate ‘what can go wrong and (occasionally) why it goes wrong, with a view to indicating the real importance of at least trying to find a solution to these problems’. The paper outlines issues with instability in problems and methods and various ways in which the user subtly does not understand what is being produced by the computer, and offers some advice for dealing with these issues both for practising scientists and those educating them, including attending an IMA conference and buying a copy of his textbook on the subject (he does apologise for the ‘plug’!).
Download the article, How to get Meaningless Answers in Scientific Computation (pdf) scanned from the 1971 IMA Bulletin, vol. 7, no. 10, pp. 296–302.
