Automatic Detection of Incomplete Instances of Structural Patterns in UML Class Diagrams

An approach for the detection of structural patterns based on UML class diagrams is presented. By using a fuzzy-like evaluation mechanism the introduced approach is able to recognize not only entire patterns but also incomplete instances. Referring to structural patterns in general the knowledge about used patterns assists a developer not only while maintaining or reverse engineering existing software, but already while designing or implementing new software. The information about the instantiation status is essential for a developer using, for example, specialization patterns that guide the extension of a particular framework. The developer can be supported by information about already instantiated patterns as well as partial instances which obviously occur rather often while developing.