Measuring ripple effect for object-oriented programs. (c2004)

Abstract

Ripple effect is a measure of structural complexity of a source code upon changing a method or a class. Ripple effect measures the amount by which this method / class may affect other methods or classes within a program, or programs within a system, if changes are made. Measurement of ripple effect has been incorporated into several software maintenance models because it shows maintainers the ramifications of any change that they may make before that change is actually implemented. Thus, computation of ripple effect provides a potentially valuable source of information. In this thesis, we propose a ripple effect measure for object-oriented programs and use it to compute an index for logical stability. Our approach is based on a new algorithm that will calculate the ripple effect for object-oriented programs at the code level by calculating both intra-class propagation and inter-class propagation for each class. It also determines the architecture ripple effect at the system level. Our method is based on matrix arithmetic for producing a ripple effect and logical stability measures and is illustrated by applying it to two examples.