Currantly, a library of all nearrings with at most 15 elements is available within SONATA. New ideas are necessary to enlarge the library beyond the current borders. This in particular includes the direct computation of isomorphism types by an incremental algorithm as well as a special consideration of the nearrings definable on the elementary abelian groups. Even with these improvements there is a natural border, simply because the number of nearrings becomes to large. Therefore, a program which randomly generates nearrings with some required properties would be highly useful for hypothesis testing.