Abstract

Scientific findings based on computer simulation evoke sceptical responses because their output does not appear to have an objective status comparable to data captured by observation or experiment. However the simulationists have been defended on grounds that their practices, like those of experimenters, carry with them their own credentials. It has been further argued that epistemic opacity is essential to the nature of computational science and that epistemology of science must cease to be anthropocentric. Such philosophical faith in software runs counter to both established practice in software engineering, and the underlying software engineering science. In computational science as elsewhere data quality can only be assured through the exercise of human critical judgment.