An example of a one-dimensional rigid set in the plane (Q1335949)

Let \(X \subset \mathbb{R}^n\) be geometrically acceptable in the sense of Borsuk \((X \in GA)\), i.e. every two points can be joined by an arc of finite length and the intrinsic metric \(\rho^*\) on \(X\) (where \(\rho^*(x,y)\) is the infimum of lengths of arcs in \(X\) joining \(x\) and \(y\)) is topologically equivalent to the Euclidean metric \(\rho\) restricted to \(X\). (Unfortunately, the author does not formulate this precisely.) The set \(X \in GA\) is rigid in \(\mathbb{R}^n\) provided that every intrinsic isometry \(f: X \to f(X) \subset \mathbb{R}^n\) (i.e. isometry with respect to the intrinsic metrics) is an isometry. The paper concerns the problem of existence of a \(k\)-dimensional set \(X\) rigid in \(\mathbb{R}^n\) for \(k < n\). The author gives a solution for \(n = 2\) and \(k = 1\). He proves that the union of all the rational lines in \(\mathbb{R}^2\) is rigid in \(\mathbb{R}^2\) (a line in \(\mathbb{R}^2\) is rational provided it has an equation \(a_0 + a_1 x_1 + a_2 x_2 = 0\) for some rational \(a_i\), \(i= 0,1,2\)). The proof is very interesting. The next step in this direction was made by \textit{Irmina Herburt} [Geom. Dedicata 49, No. 2, 221-230 (1994; Zbl 0803.54029)]. She gave a method (quite different than that of the author) of constructing \((n-1)\)-dimensional rigid subsets of \(\mathbb{R}^n\) for arbitrary \(n \geq 2\). Recently she generalized both methods and strengthened the result replacing the topological dimension by the Hausdorff dimension (the paper is prepared for publication).