An examination of materials discovery processes suggest that there can be a long lag between the creation of compounds and the discovery of their utility that would permit them to be described as materials. The goal of materials-by-design therefore is therefore dictated primarily by the ability to screen materials for function. This is the first step en route to a paradigm of dialing up the optimal material structure and composition to serve a particular function. Several issues that make even this task of screening somewhat complex. The first is that many properties of interest are not tractably calculated in a reliable way, because the underlying science is as-yet not established. The second is that materials optimization is frequently based on much more than a single performance criterion. In this talk, I will describe computational proxies that have allowed us to establish guidelines to find better phosphor materials for solid-state white lighting, better magnetocaloric materials, and some recent work on low-k dielectrics. Separately, I will describe the computational screening of all inorganic photovoltaic materials.