This thesis suggests principles that support local method development. The research is motivated by earlier findings in method research: the majority of ISD organizations which use methods develop their own variants instead of solely applying external methods as specified. Organizations, however, lack proven guidelines to develop or adapt their methods. Though method modeling languages and customizable modeling tools, such as metaCASE tools, have become available, there is a paucity of (meta)methods for method development. Our aim was to systematize a set of local method development principles that would improve the method engineering process and increase metamodeling support. Accordingly, this thesis concentrates on principles for engineering modeling techniques for ISD tools.

The ME principles developed can help organizations develop local methods and manage changes in modeling techniques. The principles are founded on extended metamodeling capabilities and on an a posteriori view of ME. In this view, experiences of method use are utilized to improve methods. The metamodeling capabilities will be summarized in Section 7.1, and the a posteriori mechanisms in Section 7.2. Finally, in Section 7.3 we propose directions for future research in this area, and complementary research topics, such as the study of implications for building tools and languages for ME.