Software metrics attempt to uncover difficult or complex components of a software system. The hypothesis is that complex components are more difficult to understand, hence they are hard to maintain and more prone to error. Discovery of these complex components can aid the software developer in selecting which components to redesign, direct the testing effort, and indicate the maintenance effort required. Previous studies have demonstrated two main concepts: (1) there exists a high correlation between design complexity and source code complexity, and (2) metrics applied to source code have a high correlation to the maintenance activity needed. This previous research motivates us to develop a methodology which uses complexity metrics throughout the software life cycle. Programmer productivity may be increased and software development cost may be reduced if error prone software is discovered early in the life cycle.