Milestoning is a theory for stochastic coarse graining in space and time and an algorithm to compute kinetics and reaction dynamics of complex molecular systems. I will discuss the theory and algorithm properties, implementation, and an application to the folding of a helix and to the recovery stroke of the protein myosin. I will review the difficulties of simulating functions of biological molecules on relevant time scales and suggest that Milestoning addresses some of these problems. Folding pathways of a helical peptide with an energy landscape that is expected to be a funnel is illustrated. The myosin example shows coupling between local molecular events and global transitions which extends to the millisecond time scale.