Local geometry of the tight 39-point octonionic code

A tight code is a point arrangement that attains a linear-programming bound on separation. Cohn, Kumar and Minton constructed such a code of 39 points in the octonionic projective plane. They conjectured that nearby tight codes, with labels and the ambient symmetries forgotten, form a 24-dimensional manifold.

Why it matters and what counts as progress

Why it matters: A manifold model would describe how optimal configurations vary near the original code. Different local dimensions or intersecting families require a description that accounts for the singularities of the configuration space.

What counts as progress: Prove or refute the proposed local dimension in the quotient by the symmetry group F4F_{4}. A classification of the original code’s local solutions or of the nearby deformation branches would give more information.

How the problem developed

  1. 23 Sept 2015 · Configuration and conjecture [1]

    Henry Cohn, Abhinav Kumar and Gregory Minton Constructed a tight 39-point code and stated the conjecture that its quotient neighborhood is a 24-dimensional manifold.

  2. 29 Oct 2026 · Local counterexample [2][3]

    Robert R. MacGregor with AI systems Showed that every neighborhood of the original code contains an open ten-dimensional Euclidean patch, which rules out the proposed 24-dimensional manifold neighborhood.

AI contribution

Sources

  1. Optimal simplices and codes in projective spaces
  2. On the local geometry of the tight 39-point code in 𝕆P2P^{2}
  3. Verification scope and AI provenance
Publication history