The history of FORTRAN I, II, and III

• Authors: John Backus
• Venue: ACM SIGPLAN Notices - Special issue: History of programming languages conference
• Year: 1978

Before 1954 almost all programming was done in machine language or assembly language. Programmers rightly regarded their work as a complex, creative art that required human inventiveness to produce an efficient program. Much of their effort was devoted to overcoming the difficulties created by the computers of that era: the lack of index registers, the lack of built- in floating point operations, restricted instruction sets (which might have AND but not OR, for example), and primitive input- output arrangements. Given the nature of computers, the services which “automatic programming” performed for the programmer were concerned with overcoming the machine's shortcomings. Thus the primary concern of some “automatic programming” systems was to allow the use of symbolic addresses and decimal numbers (e.g., the MIDAC Input Translation Program [Brown and Carr 1954]). But most of the larger “automatic. Programming” systems (with the exception of Laning and Zierler's algebraic system [Laning and Zierler 1954] and the A-2 compiler [Remington Rand 1953; Moser 1954]) simply provided a synthetic “computer” with an order code different from that of the real machine. This synthetic computer usually had floating point instructions and index registers and had improved input-output commands; it was therefore much easier to program than its real counterpart.