On Bandt's tangential distribution for self-similar measures (Q1907006)

A tangential distribution in the sense of Bandt is the weak limit of a sequence of probability measures \((M_{\eta, R})\) as \(R\downarrow 0\) on the space \(M_f(\overline B(0, 1))\) of finite measures on the closed unit ball \(\overline B(0, 1)\subset \mathbb{R}^d\). The measures \(\mu_{\eta, r}\) on \(\overline B(0, 1)\), most important for the construction, arise as image measures of a fixed self-similar measure \(\mu\) on \(\mathbb{R}^d\). The transformation behind is up to some normalizing function as in Preiss' tangent measure construction. If \(\eta\to \widehat \eta\) denotes the address to point map the transformation blows up \(\overline B(\widehat \eta, r)\) to \(\overline B(0, 1)\). The basic measure \(\kappa_R\) for \(M_{\eta, R}\) is defined on \([R, 1]\) with \(0< R< 1\) for Borel sets \(B\subseteq [R, 1]\) as \[ \kappa_R(B)= {- 1\over \log R} \int^1_R 1_B(r) {dr\over r}. \] The measure \(M_{\eta, R}\) is the image of \(\kappa_R\) w.r.t. the map \(r\to \mu_{\eta, r}\) on \([R, 1]\). The main result of the author shows in a slight more general setting that there is a unique tangential distribution \(M\) independent on \(\eta\); furthermore, an explicit integral formula is given. Introducing the notion of the average tangential measure for \(M\) an explicit formula for the order two density of Bedford and Fisher can be derived. This interesting paper closes with an open problem.