Geologic site characterization is fundamental to the selection and design of a high-level radioactive waste repository. Yet, to date, the simple question 'how many samples do we need to collect?' can not be answered for geologic site characterization. More importantly, the question 'have we collected enough samples?' can not be answered either. The objectives of this paper are broad, and their consequences extend well beyond the eight page limit of this conference. This paper attempts to address four inter-related goals. A. Explore the objectives of sampling for geologic site characterization, using the vocabulary of geostatistics. B. Motivate and describe a heuristic measure of contained information for spatially correlated geologic data using the concept of an equivalent number of uncorrelated samples. C. Review the historical record for spatially correlated sampling design with specific attention given to the sampling objectives, and D. Initiate meaningful discussion into sampling design for geologic site characterization beyond the minimum variance estimation and minimum Bayesian risk estimation of spatial averages.