Answers are grounded in this paper and cite their sources.
The maximum size of a t-intersecting code in the set of words of length n over an alphabet of size m, a problem posed by Frankl and Füredi, was determined independently by Ahlswede and Khachatrian and by Frankl and Tokushige. Extending this result to the forbidden-intersection setting, Keevash, Lifshitz, Long and Minzer proved that, for m greater than two and n sufficiently large in terms of t, (t minus one)-avoiding codes have the same maximum size and the same extremal ball constructions as t-intersecting codes. They also raised the binary case. Here, t-intersecting requires every two distinct words to agree in at least t coordinates, while (t minus one)-avoiding forbids agreement in exactly t minus one coordinates. We solve the binary case for all t at least three and n at least the square of the sum of t and one. If the forbidden Hamming distance n minus t plus one is odd, the maximum size of a (t minus one)-avoiding code strictly exceeds the maximum t-intersecting size. If it is even, the two maxima coincide, but their extremal families differ: some maximum (t minus one)-avoiding codes are not t-intersecting. We also determine all maximum avoiding codes in the even case. For odd forbidden distances, weight parity and a fixed-point-free pairing determine the maximum, while the diameter theorem gives the comparison with t-intersecting codes. For even forbidden distances, parity splitting and coordinate deletion reduce the problem to avoiding two consecutive distances. Connected components, disjoint neighbourhoods, and Harper's vertex-isoperimetric theorem yield the upper bound, which is attained by a t-intersecting construction. The diameter theorem then determines the extremal structure, and elementary binomial estimates give the stated parameter range.
No takes yet. Share an insight, caveat, or question.
Jiang et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: