An Empirical Comparison of Methods for Reengineering Procedural Software Systems to Object-Oriented Systems

摘要: This study empirically compared two methods for reengineering a procedural system to an object-oriented system. Our hypothesis was that it is possible support this process with repeatable method. The first method manual and used as baseline evaluating the second method, which based on analysis of procedure coupling. found be effective identifying objects. It produced code much smaller, more efficient, passed regression tests than Analysis metrics indicated both simpler less variability among classes. Particularly striking order magnitude difference between average cohesion metric (LCOM) methods.