This article summarizes the thinking of researchers and engineers world-wide on possible limitations of probabilistic methods in engineering, with particular emphasis upon structural mechanics. This question was posed over a ten year period (1988-1998) to numerous experts in the field. Written responses were received from 42, and they form the main body of this article. In an extensive discussion section, the author identifies some topics, trends, and thinking which seem to be prevalent in the responses, and supplements these with numerous related observations taken from the published literature, and other experts, who were not included in the original survey. This results in a references list with 52 entries. One ultimate conclusion reached is that reliability is much less known than many think it is.