find it helpful mostly in test programs to easily generate pseudo-realistic distributions of data. (A support package for koha.)