This paper presents a review and commentary on some of the fundamental facets of the life testing area of reliability. Although the emphasis employed is on statistical methods, a strictly mathematical treatment per se is avoided. Rather, an attempt is made to illustrate the fact that the problems of life testing present many complications with respect to statistical inference. It is therefore concluded that caution should be the byword in making extrapolations from a sample of life test data to a more general set of conditions. In this regard, particular emphasis is placed on the appropriate selection of life models. Finally, it is proposed that because of the statistical complications involved in many life testing problems (and other reliability problems as well) the statistician should be a member of the reliability engineering team.