ABSTRACT
This paper discusses the use of Fault Tree Analysis applied to Software. After having discussed some notational aspects, we show how a fault tree like notation can be combined with a system structure diagram to get what we call an expanded structure diagram. This diagram can be used to compute the probability of any event propagating from the system components up to the top level. It is thus not necessary to make one FT for the whole system per failure mode.
Last, the paper discusses how various computations performed on a FT, such as component criticallity, diagnostic time etc. can be applied and translated into software related topics such as the distribution of test effort and which module to check first for specific bugs.
