Mapping analytic sets onto cubes by little Lipschitz functions (Q2419684)

Call a function \(f\) on a metric space \(X\) \textit{little Lipschitz} if \[ \liminf_{r\to 0} \frac {\mathrm{diam}f(B(x,r))}{r} \] is finite for each point \(x\in X\). \textit{T. Keleti} et al. proved in [Int. Math. Res. Not. 2014, No. 2, 289--302 (2014; Zbl 1296.28008)] that if \(X\) is an analytic metric space with Hausdorff dimension greater than \(n\), then \(X\) can be mapped onto an \(n\)-dimensional cube by a Lipschitz function. The main theorem of the paper states that a perfectly parallel mapping theorem holds for little Lipschitz functions if Hausdorff dimension is replaced with the packing dimension: Theorem 5.3. Let \(X\) be an analytic metric space. If \(\mathrm{dim}_PX>n\), then there is a little Lipschitz surjective mapping \(f:X\to [0,1]^n\). The authors show that this conclusion fails if the assumption that \(X\) is analytic is dropped. Actually, they prove in Theorem 6.1 that there exist separable metric spaces with arbitrarily large packing dimension that cannot be mapped onto an interval by a continuous function. The main result requires two facts that are interesting in their own right. The first is obtained in Theorem 3.1: an analytic metric space contains, for any \(\delta >0\), a compact subset \(S\) that embeds into an ultrametric space by a Lipschitz map, and \(\mathrm{dim}_P(S) \geq \mathrm{dim}_P X-\delta\). The second is given in Theorem 4.3: a little Lipschitz function on a closed subset admits a little Lipschitz extension.