Information on nuclear properties is a vital ingredient for models of many astrophysical systems. In those astrophysical sites involving properties of nuclear few-body systems (nuclei lighter than oxygen, roughly), the data needs are somewhat different than in heavier systems. Because of the low level densities and weak Coulomb barriers of light nuclei, direct (non-resonant) mechanisms for their reactions with each other are rather important. This produces a unique set of problems in the estimation of reaction rates and the fitting of empirical rates to laboratory data. The rise of accurate quantitative models that describe light nuclear systems in terms of interacting nucleons (and, just on the horizon, QCD) in recent years provides new avenues to determine rates for use in astrophysics. Bringing together the various types of data and reaction models, and then coupling them to astrophysics, presents an array of opportunities and pitfalls. I will discuss this situation with emphasis on solar neutrinos and the big bang.