It is presently not possible to predict the structure(s)_of molecular crystals from the given structure of a single, general molecule. However, various computational methods are available to generate a set of possible crystal structures and it is likely that this set will contain the experimentally observed structure(s). The usually corrected ranking of the generated structures (with respect to the probability of occurrence) presumably is not mainly due to inaccuracies of the evaluation of energy. It appears that a full solution to the problem of molecular crystal structure prediction requires the (at least partial) inclusion of kinetic and thermodynamic effects.