Main postulates of Aleksandr Kitaigorodskii’s theory of organic crystal structures are revisited. The additive model of van der Waals atom-atom potentials giving rise to ‘bump-to-hollow’ close packing of low-symmetrical molecular bodies, suggested by Kitaigorodskii, works well for molecular crystals of non-polar hydrocarbons but requires modifications when other types of intermolecular interactions, e.g. H-bonds, have to be taken into account. Modern amount of structural data in CSD (ca. 1 mln crystal structures compared to 220 structures in the original Kitaigorodskii’s analysis) allows for several corrections and partial revision of organic molecular packing theory. In particular, ‘forbidden’ space groups are rare, but existent, in the Pareto-type rank-size distribution P(x)~x–α where P is frequency and x is the group’s rank in the list of 230 space groups. The closeness of molecular shapes and dimensions as a main driving force for solid solutions’ formation is strongly dominated by ‘like-attracts-like’ principle, first noticed for ω-disubstitued alkanes in Kitaigorodskii’s book of 1983. The problem of multiple minima of potential energy in ab initio predictions of crystal structures may be resolved looking for Gibbs potential minima (as suggested by Kitaigorodskii and Mirskaya in 1972) in disordered crystals. Finally, micro-segregation of similar molecular fragments known as hydrophobic, halogenophobic, etc. cooperative effects, points to non-additivity of weak intermolecular interactions between atoms of different elements. The main features of contemporary organic crystallography are listed and compared with that of crystallography of inorganic compounds.