This paper deals with some aspects of the use of self-checking integrated circuits in an application that manages the major risks involved in a transport system. It aims to provide an objective account of the advantages and disadvantages of this type of technology. Attention has been focused on the demands made by such integrated circuits on their environment, in particular the CPUs which control them. Nevertheless, much work still needs to be done to bring the design and testing of integrated circuits more in line with the needs of rail safety applications.