A new class of majority-logic decodable codes derived from polarity designs

Abstract

The polarity designs, introduced in [9], are combinatorial 2-designs having the same parameters as a projective geometry design $PG_s(2s,q)$ formed by the $s$-subspaces of $PG(2s,q)$, $s\ge 2$, $q=p^t$, $p$ prime. If $q=p$ is a prime, a polarity design has also the same $p$-rank as $PG_s(2s,p)$. If $q=2$, any polarity 2-design is extendable to a 3-design having the same parameters and 2-rank as an affine geometry design $AG_{s+1}(2s+1,2)$ formed by the $(s+1)$-subspaces of $AG(2s+1,2)$. It is shown in this paper that a linear code being the null space of the incidence matrix of a polarity design can correct by majority-logic decoding the same number of errors as the projective geometry code based on $PG_s(2s,q)$. In the binary case, any polarity 3-design yields a binary self-dual code with the same parameters, minimum distance, and correcting the same number of errors by majority-logic decoding as the Reed-Muller code of length $2^{2s+1}$ and order $s$.