OpenAI preprint, September 2026 · Section 1. Introduction
Corollary1.3:Complete renormings of real
,
the space of absolutely summable real sequences, with explicit norm
comparison and finite-rank approximation bounds
Equivalent norms with prescribed approximation obstructions.
Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its
declaration both in Lean and in words.
The Setting
Quantitative approximation on real concerns the operator norm needed to approximate the identity by finite-rank operators on compact sets. The real sequence space consists of the absolutely summable real sequences, with its usual norm: (1)✦
A real normed space has the bounded approximation property with bound (-BAP) if, for every compact subset and every , there are a continuous linear operator with finite-dimensional range and a real number satisfying (2)✦ The metric approximation property (MAP) is -BAP.
For an integer block parameter , the block diameter bound is (3)✦
The Claim
Theorem 1 (Quantitative renorming property). For every integer , the real vector space admits a complete norm , with its usual addition and real scalar multiplication, such that, with as in (3), (4)✦
In the paper, the result reads: for each integer , real has an equivalent norm satisfying for every , with -BAP but without -BAP, and for it fails MAP.
What Is Checked
The quantitative renorming property(4) is stated over the real sequence space(1), with its usual vector operations and the bounded approximation property defined by (2).