Bilipschitz embeddings of metric spaces into Euclidean spaces (Q1970329)

The author of the reviewing article studies the question under what conditions a metric space \((M,d)\) admits a bilipschitz embedding into some \({\mathbb R}^ n\), \(n=1,2,\ldots\). It is well known that a necessary condition for the existence of such an embedding is that \((M,d)\) be \textit{doubling} in the sense that there exists a constant \(k\) so that every ball in \(M\) can be covered by \(k\) balls of half the radius. Every \({\mathbb R}^ n\) has the doubling property and this property is inherited by spaces which admit bilipschitz embeddings into some \({\mathbb R}^ n\). A partial converse to this was given by \textit{P. Assouad} [C. R. Acad. Sci., Paris, Sér. A 288, 731--734 (1979; Zbl 0409.54020)]: If \((M,d)\) is doubling, then \((M,d^ s)\) admits a bilipschitz embedding into some \({\mathbb R}^ n\) for each \(s\) with \(0<s<1\). In general, the doubling property is not sufficient for the existence of a bilipschitz embedding into a finite-dimensional Euclidean space. In addition, the ``snowflake transform'' which replaces \((M,d)\) by \((M,d^ s)\) with \(0<s<1\), dostorts the geometry of \(M\); for example, it changes the Hausdorff dimension. Semmes is interested in transforms of metrics which do not change too much integer orders of smoothness, rectifiability of curves and surfaces, differentiability of functions, etc. Therefore, he considers embeddings of a metric space into some \({\mathbb R}^ n\) that make only modest distortions in geometry. The basic technical tool for doing this are the notions of a \textit{doubling} measure and the \textit{BPE condition} for a metric space with a doubling measure.