This paper describes the evolutionary path of hardware systems and the hardware-software interfaces that were used for GIS&T development during its “childhood”, the era from approximately the late 1960s to the mid-1980s. The article is structured using a conceptualization that developments occurred during this period in three overlapping epochs that have distinctive modes of interactivity and user control: mainframes, minicomputers and workstations. The earliest GIS&T applications were developed using expensive mainframe computer systems, usually manufactured by IBM. These mainframes typically had memory measured in kilobytes and operated in batch mode with jobs submitted using punched cards as input. Many such systems used an obscure job control language with a rigid syntax. FORTRAN was the predominant language used for GIS&T software development. Technological developments, and associated cost reductions, led to the diffusion of minicomputers and a shift away from IBM. Further developments led to the widespread adoption of single user workstations that initially used commodity processors and later switched to reduced instruction set chips. Many minicomputers and workstations ran some variant of the UNIX operating system, which substantially improved user interactivity.