Domain-specific modeling: The killer app for method engineering?

Abstract

The method creation heyday of the 1980s was characterized by convivial chaos, leading to the idea of a discipline of method engineering. Before it could grow, the unification and marketing machine of UML crushed method development into "one size fits all" design by committee in the 1990s. A scattering of brave souls went against the current, creating modeling languages specific to their own problem domain, and were rewarded with significantly higher productivity. As they seek to scale their solutions, they need help from the research world to analyze their results, and to bring to bear the learning from the early days of method engineering. © 2007 International Federation for Information Processing.