The aim of this talk is to present the problem of enumerating terms in untyped lambda calculus. The very first idea was to compute the asymptotic density of strongly normalizing closed lambda terms among all closed terms of a given size. However, the preliminary task of counting lambda terms turned out to be already non-trivial and challenging and this is therefore as the main theme. I will show recent results which were obtained due to cooperation between universities in Chambery, Versailles and Krakow.