Safety cases - What can we learn from Science?

Abstract

Safety cases typically aim to show that a system is safe through constructive argument supported by evidence: 'the system is safe' is shown to be true as a logical (or at least compelling) consequence of appropriate evidence. This paper considers an alternative point of view based on scientific method. Specifically, 'the system is safe' is viewed as a hypothesis to be challenged. The hypothesis cannot be shown to be true but trust can be built up by it surviving extensive, rigorous challenge. This approach is shown to have implications that appear to go some way towards addressing certain criticisms of safety cases raised in the Haddon-Cave report into the loss of Nimrod XV230. Some safety standards and regulatory requirements are examined from the viewpoint of 'hypothesis and challenge'. © 2011 Springer-Verlag London Limited.