Factors limiting academic participation in the development of large mathematical software packages

Abstract

The December 1975 SIGNUM meeting of Software for Partial Differential Equations seemed to formalize a trend which over the past few years has become ever more evident. Namely, the the development of medium and large scale software packages, say anything more than five hundred statements, is a project which can no longer be carried out in a university. Such projects are now almost wholly within the province of industrial and governmental laboratories. As can be seen from the program of the meeting (ACM SIGNUM Newsletter 10, 1975; No. 4, pp. 2-3), only two our of seventeen papers came from universities and one of these two papers concerned itself with a modest size algorithm for solving systems of linear algebraic equations. The reasons for this state of affairs are not hard to find. It is expensive to develop such software, and it is not altogether appropriate in a university to tie up manpower and money in such extensive undertakings. After all, departments of mechanical engineering do not build boilers, or perhaps, since new program development involves much innovation, a more accurate analogy would be to say that departments of aeronautical engineering do not build new jetliners. Also, because of tight money and excess personnel, pressures on faculty to "Produce" have increased and producing means publishing.