🤖 AI Summary
This study resolves an open problem posed by Tohăneanu: characterizing when the projective point configuration associated with a self-dual linear code fails to be arithmetically Gorenstein. For proportion-free self-dual codes, we prove that the corresponding point set is arithmetically Gorenstein if and only if the code is indecomposable—establishing an exact equivalence between a combinatorial property (code decomposability) and an algebro-geometric one (arithmetic Gorensteinness). We introduce a novel zero-one symmetrization technique to compute the dimension of the Schur square, yielding the first purely combinatorial necessary and sufficient condition for arithmetic Gorensteinness. Furthermore, we define and compute the *Gorenstein defect*, a quantitative measure of deviation from this property. All results hold over an arbitrary algebraically closed field, achieving a deep unification of combinatorial coding theory and arithmetic Gorenstein geometry.
📝 Abstract
We prove that the set of points associated to a self-dual code with no proportional columns is arithmetically Gorenstein if and only if the code is indecomposable. This answers a question asked by Toh{ă}neanu. We do so by providing a combinatorial way to compute the dimension of the Schur square of a self-dual code through a zero-one symmetrization of its generator matrix. Our approach also allows us to compute the Gorenstein defect. As a consequence, we obtain a combinatorial characterization of arithmetically Gorenstein self-associated sets of points over an algebraically closed field.