Sorting and Collating
Why did von Neumann choose to code a sorting routine for a machine intended as a high-speed scientific computer? This chapter explores the connections between von Neumann’s code and established data-processing procedures used on punched card machines. Starting in late 1944, John Mauchly interviewed several intensive users of punched cards in a variety of governmental agencies, and his notes allow us to pinpoint the moment at which the EDVAC team decided that sorting and numerical computing needed to be unified and implemented in a single high-speed machine. The early history of the mergesort algorithm and the beginnings of a systematic study of sorting procedures for electronic computers are described.