Risk management has been identified as a vitally important project management task. Hence, many risk management approaches have been developed. Unfortunately, most of them deal with the risks of a single discipline only, meaning disciplines in which either software or non-software products are developed. In contrast, nowadays many projects are highly interdisciplinary undertakings in the sense that newly developed conventional non-software products are enhanced by software. With the advent of software, many additional risks have emerged. In this paper, the differences between software and non-software products are investigated. From these differences, consequences for interdisciplinary projects are derived. It is indicated how an interdisciplinary risk management approach can cope with the consequences. An answer is given to the question of how to achieve such an interdisciplinary approach. One possible solution is presented here as an adaptation of the Failure Modes and Effects Analysis, a single discipline approach, to the needs of software. This paper is an extension of a position paper presented at the STEP2003 Workshop of Interdisciplinary Software Engineering.