Measuring Ripple Effect on Software Maintenance
Sue Black · International Conference on Software Maintenance · 1999
The Halstead and McCabe metrics give us information about the complexity of individual modules but not about modules interacting with each other. The ripple effect metric relates directly to the maintainer's task, if a change is made in one part of a system what is the likelihood of this causing a knock-on effect in a different part of the system via parameter passing or global variables? This paper gives an introduction to software maintenance models, the part that impact analysis plays within these models and a description of the ripple effect measure. Yau and Collofello's ripple effect measure seems to be a promising approach, but can be prohibitively expensive to compute. We have an approximation which appears to be acceptably accurate.