The author’s point of departure is that building today is the early architecture of the age of science. It increasingly uses scientific methods and technologies of science. Consequently there are many pressures and necessities to innovate, but resistances exist in the form of inertia of the industry, the educational deficiencies of the professions and constructors, the demanding conditions for trouble-free design and construction, and the penalties now consequent upon trouble. In order to open the way for safe innovation there has been a shift towards regulation by performance criteria in place of the former definition by specific requirements; and in order to assess performance in advance of experience, a systematic evaluation is now available. The existence of these two developments has been made possible by the growth of building science, and they in turn define the monitoring and feed-back of experience as important functions of building research for the future. There is a need and capability developing to analyse building problems with increasing precision in several directions, and the process often defines new needs for materials and techniques. This is a centreto-periphery process, and the reverse also takes place, where product makers thrust into the market innovations which result from some matching of fresh ideas to apparent needs. In all cases the needs are defined consciously or unconsciously from the context of the subsystem within which the product or component will function. Buildings are always systems comprising many subsystems. Examples are then given of directions in which the author foresees needs for new developments being defined.