These tasks generally involve the simulation of hundreds, thousands or even millions of data sets, and the subsequent estimation of a range of summary statistics from each of these. Depending on the number of data sets involved, the computational effort might be massive.
In the case of the emailer, parallelisation seemed like a fast solution to the problem, which would allow the tasked to be scaled from a desktop to a cluster environment if needs be. The below psudo-code demonstrates the use of the 'divPart' function to calculate global diversity statistics for imaginary batches of files (e.g. assume the files being analyses are separated into 10 separate folders which represent 10 separate simulation scenarios perhaps, with each folder containing 1000 genepop files), using R.