In this paper we propose the notion of material programming as a future design practice for computational composites. Material programming would be a way for the interaction designer to better explore the dynamic potential of computational materials at hand and through that familiarity be able to compose more sophisticated and complex temporal forms in their designs. The contribution of the paper is an analysis of qualities that we find a material programming practice would and should support: designs grounded in material properties and experiences, embodied programming practice, real-Time on-site explorations, and finally a reasonable level of complexity in couplings between input and output. We propose material programming knowing that the technology and materials are not entirely ready to support this practice yet, however, we are certain they will be and that the interaction design community will need to find new ways of relating to such computational materials.