This paper deals with so-called multidisciplinary design optimization (MDO), the coupling of two or more analysis disciplines with numerical optimization. MDO seems to be an important framework in applications, especially in designing algorithms. Throughout the paper all ideas and proposed research directions are explained by the concrete example of aeroelastic design. One of the motivations for the given formulation of problems is the possibility of parallel computation in nonlinear programming. Problems are discussed from a more philosophical point of view. General problems are formulated as constrained optimization problems. Some notifications are derived from linear programming, certain formulations remember the reviewer to optimal control processes. Under the assumption of differentiability linearized constraints are considered (not without misprintings!). But there is no practicable method to solve any class of problems! The authors want to realize three goals: 1. They give an overview in MDO for specialists in optimization. 2. They present an abstraction for multidisciplinary analysis design problems and a new decomposition formulation of these problems. 3. They give specialized analysis codes and introduce significant opportunities of course-grained computational parallelism.