On a construction of Burago and Zalgaller (Q1939222)

Isometric embeddings among metric spaces, or \(\mathcal C_1\) isometric embeddings among differentiable manifolds have been known for a long time. This paper tackles PL isometric embeddings, or path isometries, so we quote the definition. A map \(f:M^n\to N^q\), \(M^n\) a PL \(n\)-manifold and \(N^q\) a PL \(q\)-manifold, is called a PL isometric embedding, or path isometry iff \(f\) is a topological embedding, and length\((f(\gamma))\) = length\((\gamma)\) for all curves in \(M_n\). That a compact orientable PL 2-manifold admits PL embeddings into \(\mathbb R^3\) has been known for a long time, but \textit{Yu. D. Burago} and \textit{V. A. Zalgaller} [St. Petersbg. Math. J. 7, No. 3, 369--385 (1996); translation from Algebra Anal. 7, No. 3, 76--95 (1995; Zbl 0851.52018)] proved an analogous statement for PL isometric embeddings. Of course they needed a new construction or a method to prove this. The paper under review analyses Burago-Zalgaller's method and concludes that the method does not admit a direct extension to higher dimensions, i.e. that in any dimension \(n\geq 3\) there exists a compact PL \(n\)-manifold that can not be PL isometrically embedded into \(\mathbb R^{n+1}\) via Burago-Zalgaller's method. Besides that, the study gives a result for a class of noncompact PL 2-manifolds that admit \(\mathcal C^2\) smooth embeddings into \(\mathbb R^3\). Here is this statement. If \(M^2\) is a connected orientable PL manifold without boundary or having a finite number of compact boundary components and admits a \(\mathcal C^2\) embedding into \(\mathbb R^3\), then \(M^2\) admits a PL isometric embedding into \(\mathbb R^3\). Also, the relation between intrinsic and extrinsic curvature is examined.