Some remarks on the structure of Lipschitz-free spaces (Q1705449)

Lipschitz-free spaces \(\mathcal{F}(M)\) over a metric space \(M\) constitute a class of Banach spaces which are simple to define but quite poorly understood so far, and they presently attract the attention of a number of functional analysts, among whom the Prague school is particularly active. The present work displays several interesting results on the Lipschitz-free spaces associated with a net \(\mathcal{N}\) in a Banach space \(X\). Let me single out the following theorems: for nets \(\mathcal{N}\) in classical spaces \(X\), the space \(\mathcal{F}(\mathcal{N})\) is isomorphic to its \(l^1\)-sum. If \(X=C(K)\) where \(K\) is a metrisable compact set, the space \(\mathcal{F}(\mathcal{N})\) has a Schauder basis. It is natural to confront this theorem with the following open question: if \(M\) is a uniformly discrete metric space, does \(\mathcal{F}(M)\) enjoy Grothendieck's bounded approximation property? This problem is important. Indeed, a positive answer would show that every separable Banach space enjoys a nonlinear but uniform version of the bounded approximation property. A negative answer would provide an equivalent norm on \(l^1\) failing the metric approximation property. The paper under review contains useful tools for attacking this challenging question.