ABSTRACT
Both definitions refer to a threshold of the frequency of demand and to the ratio between demand frequency and proof-test frequency. The values considered for these two quantities seem somewhat arbitrary and, at least, neither explained nor justified. We have shown (Dutuit et al. 2008) that the factor 2 for the ratio is valid, under some assumptions, in the particular case of 1oo1 architecture. But it is not suitable for other architectures. Morover, a SIS is usually made up of three types of components (sensors / logic solver / actuators). In general, the proof-test frequencies of these components are different. So the above definitions left much too much room for interpretation.
