Preparata codes through lattices

From MaRDI portal
Publication:4544452





In this very interesting paper, the author defines an elementary family of lattices and obtains a family of extended cyclic codes with coefficients in the modular integers. This family includes as examples the Preparata codes and the Goethals codes. For certain parameters the codes are equal to extended BCH codes of length \(p^m\) and designated distance \(d\). The duals of these codes are the codes that were used by Kumar, Helleseth and Calderbank to obtain low correlation sequences. A lower bound on the Lee distance is given. The codes can be seen to be different from the BCH codes over rings, as previously defined and studied (see [\textit{I. F. Blake}, Inf. Control 29, 295--300 (1975; Zbl 0319.94003); \textit{A. R. Calderbank} and \textit{N. J. A. Sloane}, Des. Codes Cryptography 6, 21--35 (1995; Zbl 0848.94020); \textit{P. Shankar}, IEEE Trans. Inf. Theory 25, 480--483 (1979; Zbl 0418.94014); and \textit{E. Spiegel}, Inf. Control 37, 100--104 (1978; Zbl 0382.94023)]). The paper is a little bit sloppy. In Theorem 3 the author mentions leading coefficients whereas he means the coefficients \(a_1,\dots, a_{d-1}\) of \(T,\dots, T^{d-1}\). In Theorem 4 the case \(d=4\) lacks a proof (the short proof given does not exclude for instance \(d_L= 7\) from the possibilities). In Theorem 5 the defining relations for narrow sense BCH codes with designated distance \(d\) are given in a multiplicative form, however the corresponding notation (the multiplicative form) is not introduced until later on in for instance equation (11). In Theorem 7 the polynomials \(f_{e-1}, f_{e-2},\dots, f_0\) can be chosen, and the definition of the \(\rho_i\) is connected to this choice. In examples 1 and 2 the choice of \(f_{e-1}, f_{e-2},\dots, f_0\) is of course connected to the defining relations (zeroes) of the code, however this relation is not made explicit. The reader has to guess why these polynomials are chosen. Furthermore, the \(\rho_i\) are given before the \(f_{e-1}, f_{e-2},\dots, f_0\) are made explicit. Moreover in Example \(2,n\) is not specified but can be seen to be \(p^m- 1\) since it must be equal to the length of the code. Finally, the appendix is apparently added as a kind of justification and clarification of the second half of Section VI (more specifically the Theorems 8 and 9 and Examples 3 and 4). However the appendix itself is highly nontrivial and certainly not self-contained. It is incomprehensible without going back to the original source of \textit{M. Demazure} [Lectures on \(p\)-divisible groups, Lect. Notes Math. 302 (Springer) (1972; Zbl 0247.14010)].











This page was built for publication: Preparata codes through lattices

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4544452)