Demonstration code showing how (univariate) kernel density estimates are computed, at least conceptually, and allowing users to experiment with different kernels, should they so wish. NOTE: the density function in the stats package should be used for computational efficiency.