The Setting
Let be a pointed metric space. The question is how summable coefficients of point evaluations compare with the free-space norm when distinct points are at least unit distance apart and their distances to are bounded. For a real number , the separation and radius conditions are (1). (1)
The real Lipschitz space is the normed space of real functions on that vanish at and admit a Lipschitz constant .✦ Its defining condition and Lipschitz norm are (2). (2) The evaluation functional at a point is the bounded linear functional defined in (3). (3)✦ The Lipschitz-free space is the Banach space given by the closed real evaluation span in (4).✦ The Dirac embedding sends to , regarded as an element of this closed span.✦ (4) The canonical dual pairing of with is the bilinear map in (5). (5)✦
For , the Lipschitz point indicator is the function in (6), with Lipschitz constant at most .✦ The coefficient functional is the bounded linear functional obtained by pairing with . (6)
Fix an index set and a bijection , and write . The real coefficient space is the Banach space of absolutely summable scalar families, with norm as in (7). (7)✦ The coefficient analysis operator is the bounded linear map defined by (8). (8)✦
A real normed space has the bounded approximation property with norm bound , written -BAP, if for every compact subset and every there exist a bounded finite-rank linear operator and a real number satisfying (9). (9)✦
The Claim
Theorem 1. Let satisfy the separation and radius conditions (1), and let be a bijection. The Dirac synthesis operator is a Banach space isomorphism satisfying (10) and (11) for every . (10)✦ (11)✦ The space satisfies (12). (12)✦
In the paper, the result reads: if is countable, distinct points have distance at least , and its diameter is at most a finite real , then , given by , is a linear isomorphism onto with ; if , has -BAP, while if , both spaces are zero.
What Is Checked
The statement is over the closed evaluation span in under (1), with coefficients indexed through the bijection .
Evaluation and coefficient analysis satisfy the operator bounds (13). (13) The compositions of coefficient analysis and Dirac synthesis satisfy (14). (14)