The Setting
Let be a pointed metric space and let be a real normed space. Evaluation associates linear functionals to the points of . The approximation conditions concern the identity operator on and uniform error on norm-compact subsets of .
Write for the space of bounded linear operators from to , with the operator norm.
The Claim
Definition 1. The space of basepoint-vanishing real Lipschitz functions is the normed space in (1), equipped with the Lipschitz norm. (1)✦
Write for the continuous real dual of . For , the evaluation functional is the bounded linear functional given by (2). (2)✦
The real Lipschitz-free space of the pointed metric space is the closed linear span of the evaluation functionals in , with its inherited norm, as in (3). The canonical Dirac map sends to the evaluation functional , regarded as an element of this closed span.✦ (3)✦
A real normed space has the approximation property (AP) when it satisfies the finite-rank approximation condition (4). (4)✦
For , a real normed space has the -bounded approximation property (-BAP) when it satisfies (5). (5)✦
Theorem 1. For every pointed metric space and all , the canonical Dirac map satisfies the distance identity (6). (6)✦
In the paper, the result reads: for a pointed metric space , consists of real Lipschitz functions vanishing at , with the Lipschitz norm, is the closed linear span of evaluations in , and , AP means uniform approximation of the identity on every norm-compact set by bounded finite-rank linear operators, and, for , -BAP additionally bounds their norms by , while the metric approximation property (MAP) is -BAP.
What Is Checked
The definitions and distance identity are stated over arbitrary pointed metric spaces, and the approximation conditions over real normed spaces, with a real bound for -BAP.
The basepoint-vanishing Lipschitz functions form a real vector subspace of , and each evaluation has norm at most .