Abstract
Upon introducing a local change in a method or a class in an object oriented program, the ripple effect gives a measure of the amount by which this change may affect other methods/classes. That is, the ripple effect shows maintainers the ramifications of any change they may make before that change is actually implemented. In this paper, we propose techniques for measuring the ripple effect in object oriented programs and use it to compute an indicator of logical stability. First, we investigate object oriented dependences and illustrate how these dependences and some metrics can be used to determine change impact and propagation. Then, we use matrix arithmetic techniques for computing the ripple effect in object oriented programs and show their application at the intra-class, inter-class and architectural levels.