m2n: Translating Models to Natural Language Descriptions

To describe the structure of a system, the UML Class Diagram yields the means-of-choice. Therefor, the Class Diagram provides concepts like class, attribute, operation, association, generalization, aggregation, enumeration, etc. When students are introduced to this diagram, they often have to solve exercises where texts in natural language are given and they have to model the described systems. When analyzing such exercises, it becomes evident that certain kinds of phrases describing a particular concept appear again and again contextualized to the described domain. In this paper, we present an approach which allows the automatic generation of textual specifications from a given Class Diagram based on standard phrases in natural language. Besides supporting teachers in preparing exercises, such an approach is also valuable for various e-learning scenarios.