In the work presented here, we introduce a method to automatically generate a subset of a UML class model based on a user-defined criterion. The goal of this work is to allow us to automatically extract a pertinent and meaningful UML class diagram from a very large UML class model. The concept of model slicing is introduced as a means to support maintenance through the understanding, querying, and analysis of large UML models. The specific models being examined are class models as defined in the Unified Modeling Language (UML). Model slicing is analogous to classical program slicing. Since UML class models do not explicitly embody any behavioral aspect by themselves, models slices are computed in a context-free manner. The paper defines and formalizes the concept of context-free model slicing. A concrete application of model slicing in software maintenance is presented to support the usefulness and validity of the method. In the next section (2) we describe and formally define our appr...
Huzefa H. Kagdi, Jonathan I. Maletic, Andrew Sutto