The following argument is presented: The actual achievements of computational fluid mechanics, great though they are, scarcely measure ~p to the power of the numerical schemes and of the computers that are now available. In practice, the time and energies of would-be users of computational schemes are often spent on the devising and maintaining of the computer programs which embody these schemes; and Uhese programs, being vulnerable to damage arising from well-meant al teratlons, seldom perform well for long periods, once they have left their originators' hands.
Progress in the application of numerical models, and in the refinement of the physical models, and in the refinement of the physical hypotheses which they embody, could be greatly accelerated if practitioners were enabled to employ a reliable and economical standard computer programl incorporating the well-established balance laws of fluid mechanics, heat transfer and mass transfer, and structured so that particular material properties, geometries and boundary conditions could be easily incorporated. Such a program, if continuously and centrally maintained, could free the computer-modelling fraternity of the chores of code construction and maintenance; it would thus permit them to apply their innovative talents to the development and application of improved physical hypotheses.
The PHOENICS code system has been developed in order tO meet this need. It9 naturel current capabilities and development possibilities will be described in the paper.
Because every student learns computer program ].