This paper gives an exposition of the convention, introduced by D. van Dantzig in 1947, of indicating randomness by underlining the relevant symbols. This method of distinguishing between random variables (or elements) and non-random ones serves a multiple purpose. It saves symbols and may be applied to symbols which have been introduced first in a non-random quality. It clearly marks the line between measure theory and probability theory and, in statistics, between descriptive and stochastic models. This underlining notation has proved to be very useful during its nearly twenty years of use in Holland and many Dutch authors use it in their publications. The present description of the rules of this convention is meant to facilitate the reading of these papers and to deepen the understanding of the merits of this notation.
Abstract