All of genetics is concerned with associating specific trait alleles with specific phenotypes. The challenge is to find the trait alleles in a vast genome given a phenotype. Whether one chooses linkage or association studies for tracking such alleles, the essential problem is detection of those segments of the genome shared identical-by-descent (IBD) between a group of affecteds. The current discussions, and sometimes disagreements, of the relative utility of each method depends on our elucidating the nature of human genetic variation. I shall describe empirical observations of sequence variation in genes and genomic segments that provide insights into measuring IBD segments in unrelateds, and thus how association studies can be designed.