Emergent laws of method and class stereotypes in object oriented software

The paper investigates concepts of method and class stereotypes, focusing on understanding OO design abstraction at a lower level than design patterns. Stereotypes are powerful semantic mechanisms and represent generalizations that reflect an intrinsic or atomic behavior of a method or a class. First, we present a mechanism to automatically reverse engineer these stereotypes from existing systems along with a means to re-document methods and classes with their corresponding stereotypes. Second, based on the distribution of method stereotypes we propose techniques for automatic identification of class stereotypes, systems classification, and the characterization of changes during software evolution. We anticipate that these new techniques will better support program comprehension and design recovery, and could be used to build smarter reverse engineering tools. The results of this paper are based on the author's doctoral dissertation.