Scientific scale of reliability

Reliability Theory, since it\´s beginnings in 1950s, has been based on mathematical theorems rather then on scientific theories. Massive attempts where made to further applications of the existing mathematical and statistical methods and analysis without attempts for understanding "failure mechanics". Then, in 1980s, practicing reliability engineers and analysts, who have neither ability nor need to understand the mathematics, turned to what they have had, which is enormous practical experience of the observed failure modes of existing systems. Thus, a large number of "practical reliability methods" have been developed and used, all of which were based on the failure mode, effect and criticality analysis, but still without understanding and addressing failure mechanics. Consequently, during the last 50 years the Reliability Theory made very little progress, a part from a few exceptions, in the direction of becoming the science, in terms of making accurate predictions that could be confirmed with practical observations. The reason is very simple; neither statistics, which does not study causes of statistical behaviour, nor engineers whose "applied methods" were focused on meeting contractual and legal requirements, were able to provide a fertile ground for the development of reliability.