This chapter charts the effects of programming languages, workstation availability, user demands, distributed computing, and graphical capabilities on the use and evolution of mathematical software from the 1940s to the present day. The ideas of John Rice, whose visionary ideas and key contributions have often cut a path for others to follow, are intertwined throughout this development.
Mass-market mathematical software has become profitable due to the demands of nonprogrammers to solve mathematical problems. However, it is no longer the case that the programmers of the numerical cores of these packages are experts in numerical analysis or that the end users have the necessary expertise to recognize difficult problems or poor results. The author’s peek into the future provides a ray of hope for everyone; with a bit of luck, scientists and engineers will have easy access to top quality software, and both numerical analysts and mathematical software engineers will still have jobs.
This paper is not just biographical; rather, it provides an excellent overview of an important and complex field that impinges on many aspects of computer science. As such, this chapter should be enforced reading for all new (and a good many old) users and producers of mathematical software.