Discovery in the context of application

This article addresses several issues raised by a very particular pattern of knowledge generation, called ''discovery in the context of application.'' The case of a hypersonic aircraft research program illustrates that the organizational design of such programs is a very complex matter. The problem cannot be dealt with in the conventional way by having different organizational forms for different phases of the process-say, one for the research phase another for the development phase, a further for the testing phase, and so on-because of its integral nature. Two modes of organization are considered but neither, on its own, seems to be quire adequate to govern the management of the whole process. The need is to deal with the problem of finding a balance between two (distinct) organizational imperatives: the need to manage the interaction between a large number of projects that are carried out in different locations, and the need to bring a wide range of problem solvers into a single entity. The former mode is necessary to facilitate experiments in different directions and to decrease the risk of missing a possible successful design, the latter to develop the infratechnologies and instrumentalities that will underpin the research agenda. (C) 1996 Elsevier Science Inc.