In this paper I examine the issue of whether we can trust software systems and put forward an argument that in an absolute sense the answer must necessarily be no for a multitude of reasons. The paper then examines the question of whether this is an actual impediment to the successful application of software with particular reference to automotive applications. In particular I examine the question of whether our expectations of software are compatible with the realities of road vehicle manufacture and use. I conclude that with good methodology and integration within a whole vehicle development process, software based systems can, and will provide levels of safety above those which are experienced today so long as certain critical constraints are met. .