On the De Boer-Pellikaan method for computing minimum distance (Q992803)

M. De Boer and R. Pellikaan gave a method to compute the minimum distance \(d\) of a linear error correcting code \(\mathcal{C}[n,k,d]\) defined over a (finite) field \(K\), in [\textit{A. M. Cohen, H. Cuypers, H. Sterk}, Some Tapas of Computer Algebra. Algorithms and Computation in Mathematics. 4. Berlin: Springer. (1999; Zbl 0924.13021)]. The present paper discusses some consequences and applications of that method. Section 1 summarizes the De Boer-Pellikaan method, which implies certain ideals \(I_j,\,\, j=1, \dots, n\),\, generated by linear forms in \(K[x_1,\dots, x_k]\), forms determined by the columns of a generating matrix of \(\mathcal{C}\). Section 2 derives some results concerning the (projective) codewords of minimum weight \(d\)\, of \(\mathcal{C}\) and gives a necessary and sufficient condition in terms of \(I_{d+1}\)\, for \(\mathcal{C}\) to be maximum distance separable. Section 3 presents the main result of the paper (Theorem 3.1), which proves that for all \(j=1,\dots , d\),\, \(J_j=m^j\),\, where \(m=(x_1,\dots , x_k)\)\, maximal ideal of \(K[x_1,\dots, x_k]\) and Section 4 derives some consequences for the minimal codewords of \(\mathcal{C}\). Finally Section 5 compares the De Boer-Pellikaan method with the alternative method of \textit{J. Migliore} and \textit{C. Peterson} to find \(d\)\, using Chow forms [J. Symbolic Comput. 37, 403-413 (2004; Zbl 1137.14316)]. The paper gives three reasons that do the De Boer-Pellikaan method \`\` more useful in practice''.