A dichotomy for the Weierstrass-type functions (Q2052223)

The authors consider Weierstrass-type functions \[ W(x)=\sum_{n=0}^{\infty}\lambda^{n}\phi(b^{n}x), \] where \(\phi:\mathbb{R}\to\mathbb{R}\) is a real analytic periodic function, \(\lambda\in(1/b,1)\) and \(b\ge2\) is an integer. They prove that either \(W(x)\) is real analytic, or the Hausdorff dimension of its graph is equal to \(2 + \log_{b}(\lambda)\). Furthermore if \(\phi\) is not constant the first alternative only happens for finitely many \(\lambda\). This is an impressing result on the geometry Weierstrass-type functions.