| As the software industry has matured, we have shifted our resources from being devoted to developing new software systems to making modifications in evolving software systems. A major problem for developers in an evolutionary environment is that seemingly small changes can ripple throughout the system to cause major unintended impacts elsewhere. As such, software developers need mechanisms to understand how a change to a software system will impact the rest of the system. Although the effects of changes in object-oriented software can be restricted, they are also more subtle and more difficult to detect. Maintaining the current object-oriented systems is more of an art, similar to where we were 15 years ago with procedural systems, than an engineering skill. We are beginning to see "legacy" object-oriented systems in industry. A difficult problem is how to maintain these objects in large, complex systems. Although objects are more easily identified and packaged, features such as encapsulation, inheritance, aggregation, polymorphism and dynamic binding can make the ripple effects of object-oriented systems far more difficult to control than in procedural systems. The research presented here addresses the problems of change impact analysis for object-oriented software. Major results of this research include a set of object-oriented data dependency graphs, a set of algorithms that allow software developers to evaluate proposed changes on object-oriented software, a set of object-oriented change impact metrics to evaluate the change impact quantitatively, and a prototype tool (ChaT) to evaluate the algorithms. This research also results in efficient regression testing by helping testers decide what classes and methods need to be retested, and in supporting cost estimation and schedule planning. |
