Fundamental Postulate in Structural Safety

Arguments are given in favor of the point of view that structural safety should be quantitatively expressed in terms of a pair of numbers, ($p_{\mathrm{th}}\text{,}$ $p_{\mathrm{gr}}$), rather than by a single number. In this pair, $p_{\mathrm{th}}$ is the theoretical failure probability as it results from the structural reliability theory disregarding the existence of gross errors. The second number, $p_{\mathrm{gr}}\text{,}$ is some measure of proneness to failure due to gross errors. It is postulated that structural optimization is a two variable problem with $p_{\mathrm{th}}\text{,}$ and $p_{\mathrm{gr}}$ as independent variables. The paper is a reaction to recent suggestions of merging reliability theory results given in terms of probabilities with judgements concerning gross errors expressed, for example, by use of the concept of fuzzy sets.