On boundaries on the predual of the Lorentz sequence space (Q2644028): Difference between revisions

For a complex Banach space \(X\), let \(A_u(B_X)\) denote the uniform algebra of all uniformly continuous functions on the closed unit ball \(B_X\) which are analytic in the open unit ball, endowed with the sup norm. A subset \(S\) of \(B_X\) is called a boundary if the supremum on \(S\) yields the sup norm. \textit{J.\,Globevnik} [Math.\ Proc.\ Camb.\ Philos.\ Soc.\ 85, 291--303 (1979; Zbl 0395.46040)] described those boundaries for \(X=c_0\) and proved that there is no such minimal boundary. This is a phenomenon which cannot happen if \(X\) is a finite-dimensional space. Subsequent research on the topic has been done by several authors. In the paper under review, the authors continue such research for \(X\) the canonical predual of a Lorentz sequence space. They provide characterizations of boundary sets \(S\). One of them (Theorem~3.2) is in terms of the vanishing distance from \(S\) to every strong peak set. Another one (Theorem~3.1) is built on the containment of a suitable sequence of finite-dimensional projections of \(S\) to some sort of tori, which the authors introduce in a similar way as Globevnik [loc.\,cit.]. Remarkably again, they show that no such minimal boundary exists in the present case.