"At present, the generation of technology is decisively determined by a close interaction of academic und industrial research. Conventional linear models of the innovation process reflect this interaction in an insufficient way. More recent approaches with recursive models reflect the innovation process more adequately, but only partly help to achieve a clearer understanding, as they do not permit definite linkages to the involved actors and a description of the development in time. The model suggested in this paper explicitly introduces the actors as well as the time dimension and clearly differentiates between innovation stages and types of research. The model is illustrated by the example of neural networks, an area with close communication structures between academic and industrial researchers which can be described as technoscientific community." Erschienen in: Soziale Welt 47. Jahrg., H. 2 (1996), pp. 250-265