In his synthesis on industrial dynamics, Malerba called for a renewal of the models for the dynamic analysis of innovation and the evolution of industries [1]. To go this way we investigate the relationship between knowledge dynamics, innovation dynamics, and sectoral growth in the particular case of Schumpeterian "development" [2]. Our analysis is based on a model where economic actors (suppliers and customers) are represented by design functions, endogenizing the generation of "unknown" products, the regeneration of competences and of utility functions. We use the model to simulate four situations of industrial dynamic characterized by the (successful or impeded) emergence of novelty: automotive industry, pharmaceutical and biotech industry, semiconductor industry and orphan innovation in cleantech. This model shows that the success of "novelty-oriented" industrial dynamics depends on the efficiency of the coupling between design functions in the economy. We show that 1) good suppliers' profit and customers user-value relies on a sparing of knowledge and novelty; 2) coupling is based less on the initial level of competences and knowledge capitalization than on learning from "unknown" products; 3) learning from the unknown creates externalities, so that the exploration of the unknown appears as a new kind of "common good".