True dimension of some binary quadratic trace Goppa codes (Q5947348): Difference between revisions

Let \(g(z)\in \mathbb{F}_{q^m}[z]\), \(L=\{\alpha_1, \dots, \alpha_n\} \subset \mathbb{F}_{q^m}\) such that forall \(i\), \(g(\alpha_i)\neq 0\). The Goppa code \(\Gamma (L,g)\), of length \(n\) over \(\mathbb{F}_q\), is the set of codewords, i.e., \(n\)-tuples \((c_1, \dots,c_n) \in\mathbb{F}_q^n\), satisfying \[ \sum^n_{i=1} {c_i\over z-\alpha_i} \equiv 0\pmod {g(z)}. \] The dimension \(k\) of \(\Gamma(L,g)\) and its minimal distance \(d\) satisfy \(k\geq n-m \deg g(z)\) and \(d\geq\deg g(z)+1\). It is well known that it is a hard problem to compute the true dimension (and minimal distance) of any Goppa code. A lot of work has been done on special classes of Goppa codes in order to improve the general bound on the dimension. Let \(a(z)\) and \(b(z)\) be two arbitrary elements of \(\mathbb{F}_{p^{ms}} [z]\). A trace Goppa code is a \(\Gamma(L,g)\) code where \(g(z)=a(z) \text{Tr}_{\mathbb{F}_{p^{ms}}: \mathbb{F}_{p^s}} (b(z))\) and \(L=\mathbb{F}_{p^{ms}} \setminus \{z\in \mathbb{F}_{p^{ms}}\), \(g(z)=0\}\). Depending on the value of \(p\) and \(m\), three new bounds are given in [\textit{P. Véron}, IEEE Trans. Inf. Theory 44, 290-295 (1998; Zbl 0902.94020)] for the dimension of such codes. Moreover it is proved that these codes never reach the general known bound. The aim of this paper is to prove that the true dimension of binary Goppa codes defined by \(g(z)=\text{Tr}_{\mathbb{F}_{2^{2s}}: \mathbb{F}_{2^s}}(z)\) is \(n-2s\deg g(z)+3s-1\), proving this way a conjecture stated in [\textit{A. M. Roseiro}, \textit{J. I. Hall}, \textit{J. E. Hadney} and \textit{M. Siegel}, IEEE Trans. Inf. Theory. 38, 1130-1133 (1992; Zbl 0749.94013) and \textit{P. Véron} (loc. cit.)].