With the advent of massively parallel computer systems it is important to reconsider the algorithms and implementations of quantum chemical methods.In my talk I will discuss the merits of subsystem approaches that allow for optimal use of parallel computers and offer a chemically transparent interpretation of the computed data.