“…A first and remarkable observation which was already made as far back as the 1940's by Ulam and von Neumann [30] is that the parameter 2 can be thought of as "stochastic" in a precise mathematical sense (more formally, the corresponding map f a admits an ergodic invariant probability measure which is absolutely continuous with respect to Lebesgue measure). In 1981, Jakobson [6] and then in 1985 Benedicks and Carleson [2], showed that the set Ω + ⊂ Ω of such stochastic parameters is actually "large" in the sense that it has positive Lebesgue measure (see also some generalizations in [24,28,23,20,27,12,13]). Interestingly, however, it is also "small" in the sense that it is topologically nowhere dense, a fact that follows from the remarkable result that the set Ω − ⊂ Ω of "regular" parameters (for which almost every initial condition is eventually attracted to a periodic orbit) is open and dense in Ω [4,15,16] (see also generalizations in [10,11]).…”