Quantitative measurement of program quality

A black-box approach typifies current software quality assurance procedures: a program is good it it satisfies certain operating specifications. While it is common to manage the development of software under quality assurance systems previously devised for hardware, the tools of measurement are not transferable owing to the very basic differences in the nature of hardware and software. In the absence of specific, applicable quantitative measurement tools there exists no means of defining the desired level of quality in a computer program, where quality is considered as something beyond correct program functioning, nor of ascertaining whether the desired level has been achieved. A user should be able to specify precisely how good a product he wishes to buy, such things as how easy the program should be to run production with and how easily it can be modified. Rarely can the user even discuss these factors, much less specify the extent of their importance to him.