Excerpted from Systems Architecting, Creating and Building Complex Systems, Eberhardt Rechtin, Prentice Hall, 1991.
SCIENTIFIC RESPONSE: BETTER MEASUREMENT TECHNIQUES
A common occurrence in failure investigations is a demand for new diagnostic methods to determine the underlying physical causes of failures not yet explained. The need is to convert so-called random failures into well-determined controllable events. Such has been the case for the treatment of diseases, for Titan rocket motor failures, spacecraft electrical discharge short circuits, radiation-damaged integrated circuits, faulty circuit boards, and the collapse of structures.
In contrast with redundancy and other problem-expansion approaches to increased quality, this one attempts to reduce large-scale failures to smaller, even subatomic, phenomena that can be eliminated by design or process control. The resulting instrumentation now ranges over almost all the electromagnetic and sonic spectrums, heavily supported by data processing, simulation, models, and expert systems.
Better measurement capability is valuable in other important ways. It pushes still higher the level of ultraquality than can be achieved with some confidence. It extends further the point where ultraquality's "beyond measure’’ begins. And it relieves to some extent the need to carry out all the other responses to this challenge.
You might also be interested in these articles...
Software Architecture as Code