A quantitative isoperimetric inequality on the sphere (Q2360365)

In the very interesting paper under review, the authors prove a quantitative version of the isoperimetric inequality on the sphere with a constant that is independent of the volume of the set \(E.\) More precisely, let \(S^n\) be the \(n\)-dimensional sphere with \(n\geq2,\) and \(E\subset S^n\) a smooth measurable set having the same measure as a geodesic ball \(B_\theta\subset S^n\) for some radius \(\theta\in(0,\pi).\) Setting \(\mathbf{P}(E)\) for the perimeter of \(E,\) that is, \(\mathbf{P}(E)=\mathcal{H}^{n-1}(\partial E),\) a classical result of \textit{E. Schmidt} [Math. Z. 49, 1--109 (1943; Zbl 0028.31303)] asserts that \[ \mathbf{P}(E)\geq \mathbf{P}(B_\theta), \] with equality reached if and only if \(E\) is a geodesic ball. The authors derive a fine quantitative version of that inequality by proving that there exists a constant \(c(n)\) such that \[ \mathbf{P}(E)-\mathbf{P}(B_\theta)\geq c(n) \boldsymbol{\beta}^2(E) \] for any set \(E\subset S^n\) of finite perimeter and volume \(|E|=|B_\theta|\) for some \(\theta\in(0,\pi).\) Here the quantity \(\boldsymbol{\beta}(E)\) is refereed as the \textit{oscillation index} of the set \(E\) that measures the distance between the distributional derivatives of \(\chi_E\) and of \(\chi_{B_\theta},\) and is defined by \[ \boldsymbol{\beta}^2(E)=\dfrac{1}{2} \min_{p_0\in S^n} \int_{\partial E} \big|\nu_E(x)-\nu_{B_\theta(x)(p_0)}(x)\big|^2d\mathcal{H}^{n-1}(x), \] where \(\nu_E(x)\) is the outward normal to \(\partial E\) at \(x\in \partial E,\) lying in the tangent plane to \(S^n\) at \(x,\) and \(\nu_{B_\theta(x)(p_0)}(x)\) is the outward normal to the geodesic ball \(B_\theta(x)(p_0)\) centered at \(p_0\) whose boundary passes through \(x.\) The difficulties in deriving the above isoperimetric inequality arise from the fact that the sphere does not admit natural intrinsic scalings, which could allow reduction to the Euclidean case. The authors avoid this obstacle by a subtle rewriting the oscillation index as \[ \boldsymbol{\beta}^2(E)=\mathbf{P}(E)- (n-1) \max_{p_0\in S^n} \int_{E} \dfrac{x\cdot p_0}{\sqrt{1-(x\cdot p_0)^2}} d\mathcal{H}^{n}, \] and then estimating the singular integral \[ \int_{E} \dfrac{x\cdot p_0}{\sqrt{1-(x\cdot p_0)^2}} d\mathcal{H}^{n} \] and its maximum with respect to \(p_0,\) independent of the volume of \(E.\)