If MDE is the Solution, then What is the Problem?

For nearly ten years, modern forms of software modeling have been used in various contexts, with good apparent success. This is a convenient time to reflect on what has been achieved, where we stand now, and where we are leading to with Model-Driven Engineering (MDE). If there is apparently some consensual agreement on the core mechanisms, it is much more difficult to delimitate the scope and applicability of MDE. The three main questions we have to answer in sequence are: 1. What is a model? 2. Where are models coming from? 3. What may models be useful for? There is now some consensus in the community about the answer to the first question. A (terminal) model is a graph conforming to another graph usually called its metamodel, and this terminal model represents a system. Terminal models and their metamodels are similarly organized and may be unified as abstract models, yielding a regular organization. In such an organization, some of the models (e.g., a transformation) may be executable. The relation of conformance between a terminal model and its metamodel provides most of the information on the first question.