# A Projection Decoding of a Binary Extremal Self-Dual Code of Length $40$

## TL;DR

This paper introduces two efficient decoding algorithms for a specific extremal binary self-dual code of length 40, utilizing projections onto a Hermitian self-dual code over GF(4) to simplify decoding.

## Contribution

The paper presents novel projection-based decoding algorithms for a binary extremal self-dual code, reducing complexity by leveraging a Hermitian self-dual code over GF(4).

## Key findings

- Two new decoding algorithms for the [40,20,8] code.
- Algorithms are implementable by hand, unlike syndrome decoding.
- Decoding complexity is reduced through projection onto a smaller code.

## Abstract

As far as we know, there is no decoding algorithm of any binary self-dual $[40, 20, 8]$ code except for the syndrome decoding applied to the code directly. This syndrome decoding for a binary self-dual $[40,20,8]$ code is not efficient in the sense that it cannot be done by hand due to a large syndrome table. The purpose of this paper is to give two new efficient decoding algorithms for an extremal binary doubly-even self-dual $[40,20, 8]$ code $C_{40,1}^{DE}$ by hand with the help of a Hermitian self-dual $[10,5,4]$ code $E_{10}$ over $GF(4)$. The main idea of this decoding is to project codewords of $C_{40,1}^{DE}$ onto $E_{10}$ so that it reduces the complexity of the decoding of $C_{40,1}^{DE}$. The first algorithm is called the representation decoding algorithm. It is based on the pattern of codewords of $E_{10}$. Using certain automorphisms of $E_{10}$, we show that only eight types of codewords of $E_{10}$ can produce all the codewords of $E_{10}$. The second algorithm is called the syndrome decoding algorithm based on $E_{10}$. It first solves the syndrome equation in $E_{10}$ and finds a corresponding binary codeword of $C_{40,1}^{DE}$.