Decomposition properties of John domains in normed vector spaces (Q662061)

Let \(E\) be a real normed vector space with \(\dim E\geq 2\). A domain \(D\) in \(E\) is said to be a \(c\)-John domain if each pair of points \(z_1,z_2 \in D\) can be joined by a rectifiable arc \(\alpha\) in \(D\) such that \[ \displaystyle \min_{j = 1,2} j(\alpha [ z_j, z]) \leq c \cdot d (z, \partial D) \] for all \(z \in \alpha\), where \(\alpha [z_j, z]\) is the subarc of \(\alpha\) between \(z_j\) and \(z\), \(l (\alpha [z_j,z])\) the length of \(\alpha[z_j,z]\), and \(d(z, \partial D)\) the distance from \(z\) to the boundary \(\partial D\) of \(D\). In general, \(D\) is called a John domain if \(D\) is a \(c\)-John domain for some \(c > 0\). A domain \(D\) in \(E\) is said to have the John domain decomposition property if there exists a constant \(c\) with the following property: for each pair of points \(z_1,z_2 \in D\), there exists a subdomain \(D_0\) of \(D\), such that \(z_1,z_2 \in \overline{D_0}\) and \(D_0\) is a simply connected \(c\)-John domain. In the article under review, the authors prove that a domain \(D\) in \(E\) is a John domain if and only if it has the John decomposition property. It is a generalization and improvement of the result for uniform domains by \textit{F. W. Gehring} and \textit{B. G. Osgood} [J. Anal. Math. 36, 50--74 (1979; Zbl 0449.30012)].