Real life law cases can be analyzed and modeled, but the reverse direction is possible, too: Creating artificial cases based on an ontology. Combined with adaptivity, such cases can be used in E-Learning as exercises for learners. This paper presents an approach for generating cases based on assembling text fragments combined with an ontology for verifying whether these fragments are compatible and the resulting case is factually and legally possible. An exemplary implementation of the subject area of domain name disputes is shown as a proof of concept.