It is a difficult and not entirely rewarding task to attempt to foretell the future. On the one hand, there are many voices saying the same thing and one runs the risk of repeating the trite and obvious; on the other hand, there are a few voices describing developments that might be very exciting but have a low probability of realization. My inclination is to side with the majority, not so much with an aim to look back after a few years with some satisfaction, but to indicate where in all probability current trends in computers are leading and what their impact might be on computational fluid mechanics. In addition, my perspectives are those of one who comes from a background of engineering, teaching, large-scale numerical modeling , and, currently, the management of a computing facility devoted to solving the mos� challenging computational problems in the atmospheric sciences.Section 2 looks first at the current healthy status of computational fluid mechanics, particularly in the atmospheric sciences, and indicates the need for considerable additional computing resourccs in the years ahead.Section 3 describcs the substantive changes in the design and development of largc-scale computational systems that are currently under way. These changes are causcd by both technological and economical factors. Section 4 then describes the trends in technology and the economics of those trends.Section 5 examines the growth of data processing in fluid dynamics and the establishment of large data bases for both computational models and experiments. Section 6 argues that the generation and sharing of software for applications in fluid mechanics may be an even more important issue in the 1980s than the design and development of hardware.Finally, the mixed prospects for computational fluid mechanics are summarized in Section 7.