Requirements concerning nuclear calculation methods and how they are to be met by a common R+D programme

For the design of reactor cores the neutron physicist will require mainly three ''instruments'': Computing programmes; Core and materials data; Specialist know-how. The requirements to be met by these ''instruments'' can be subdivided into three complexes: Accuracy; reliability; efficiency. In the early seventies, manufacturers' requirements concerning accuracy were for the first time systematically compiled under the DeBeNe programme and compared with failure rates known at the time. This lead to relevant criteria for setting up R + D schedules for successive years. The term ''reliability'' means safety against systematical or sporadically occurring failures. Redundant research, routine in handling computing programmes and quality assurance thereof are some of the relevant aspects in this context. Efficiency of the methods does not only or mainly refer to the computing time required by the programmes. Simplicity of handling and avoiding superfluous manwork, in particular in calculating series of programmes of the same kind is more important. Such influences of rationalization are most important for the manufacturer. This paper shows to what extent the requirements described were met by the past R + D programme. It also indicates further research tasks that still need to be fulfilled. It finally deals with the question whether, and if so, to what extent cooperation of various partners may lead to more '' harmony'' in the use of identical procedures. (orig.)