Give the distances between every data from the first input set and the data from the second input set. It outputs the size distribution. Each point (x, y) tells you that there exists y pairs of elements which are separated by x nucleotides.
The general algorithm is the following. For each element of the first input set, it finds the closest element of the second set and computes the distance between the two elements. The distance is zero if the two elements overlap. This distance may not exist if the element of the first input set is alone on its chromosome (or contig).
Actually, considering an element from the first input set, the algorithm will look at the vicinity of this element (1kb by default). You can increase the size of the vicinity using the appropriate option.
As in compare overlapping, you can shrink or extend your sets of genomic coordinates, so that you can get the distance between starts of reads and starts or genes, for instance. You can also compute the distance from elements which are on the same strand only (which is not the case by default) or on the opposite strand only.
You have several options for the output plot. You can first choose the region on the x-axis you want to plot. You can also display histograms instead of line plot. In this case, the data are summed into buckets, whose sizes are given as an option. For instance, a bucket of size s at the point (x, y) means that there are y pairs of elements which are separated by x to x + s nucleotides.