Reliability and its relation to suitability and predictability

Reliability, like a great many other words, means different things to different people. Let me illustrate with a purely imaginary example. Suppose we have a small vacuum tube with amplification, mutual conductance, and plate impedance all of useful magnitudes. Tubes of this type have been made in large quantity. Their characteristics, when measured at the factory have very good uniformity---all are within ±1 per cent of their nominal value. Every tube that has been examined has kept its characteristics within these narrow limits throughout its entire life, and a large fraction of the tubes made have been thus examined. These characteristics are entirely independent of such things as ambient temperature and mechanical shock.