The problem of the optimal synthesis of chemical units, in the sense of a given technical or economic criterion, is one of the oldest in the field of process engineering. As early as the late 1950s, its formulation was determined. Since then, the field has seen the publication of several hundred articles on the subject with eight main overall reviews (Hendry et al. (1973), Hlavacek (1978), Westerberg (1980), Stephanopoulos (1981), Nishida et al. (1981), Umeda (1983), Westerberg (1985), and, most recently, Floquet et al. (1988). Not a year has passed without the announcement of new promising methods that are then applied to a number of examples found in the literature. Our article is no exception. In it, an indication is given of the avenues that we consider to be worth exploring for the integration, by the industrial user, of the results of almost 40 years of academic research. Our faith is pinned on two factors: (1) The formidable increase in computer power capability that has opened up the possibility of tackling optimal process synthesis using hypotheses that could not have been handled before. (2) The better integration of “human factor”, that is to say industrial experience, within the process engineering sectors. These two considerations combined - the introduction of new tools and the intelligent integration of “old tricks” - leads us to believe that it is now possible to envisage the actual optimal synthesis of separation sequences and, more generally, industrial units themselves; groups of interconnected reactors and separators, complete processes.