
In June 1962, the first American space probe to Venus (Mariner I) went off course and had to be destroyed because of an error in one of the guidance programs in its onboard computer.
One statement of the program, though syntactically correct, had a meaning altogether different from that intended by the programmer. Although few bugs have such spectacular effects, errors in computer programs are frequent and influential. There has been substantial effort recently to apply mathematical rigor to the programming process and to enable the accuracy of the machine to compensate for the error-prone human mind.
In the late nineteenth and early twentieth century, mathematics underwent a process of formalization and axiomatization, partially in an effort to escape from paradoxes and logical errors encountered by previous generations of mathematicians. A similar process is underway in the development of a logical theory of programs.’ This theory has already made our understanding of programs more precise and may soon facilitate our construction of computer programs as well.
Logical techniques are being developed
- to prove programs correct,
- to detect programming errors,
- to improve the efficiency of program operation,
- to extend or modify existing programs, and even
- to construct new programs satisfying a given specification;
In the last decade, this field of research has been extremely active; it now has the potential to exert a deep influence on the way computer programs are produced.