The dimension of subcode-subfields of shortened generalized Reed-Solomon codes (Q356722)

The main goal of the paper is to provide a way to generate cyclic codes with performances at least as that of the best codes listed in [\textit{M. L. Grassl}, ``Bounds on the minimum distance linear codes'', \url{http://www.codetable.de}]. The codes are of type subfield-subcodes (SFSC), covered by Generalized Reed-Solomon codes (GRS) which offer good results in coding and decoding procedures. Based on the ideas from [\textit{A. M. Roseiro} et al., IEEE Trans. Inf. Theory 38, No. 3, 1130--1133 (1992; Zbl 0749.94013)] -- to use the kernel of the associated trace map -- the authors obtain a new formula for a lower bound for the minimum distance of a specific type of SFCS codes. Two algorithms for generating new codes are proposed. Some results (very performant codes over \(\mathbb{F}_2, \mathbb{F}_3\) and \(\mathbb{F}_5\)) obtained following Algorithm 1 are listed. From these codes other good codes result using puncturing and shortening procedures usually used in practical implementation of Reed-Solomon codes. The paper is interesting and valuable by its new way of generating the new type of SFSC codes. But I did not like the lack of rigour in writing the paper. For example, many definitions are omitted: see \(f\) in Definition 1, \(g'\) in Theorem 2.6 etc (to understand some notations the reader must know the domain); two notations (\(|X|\) and \(\#X\)) are used to describe the same thing (the cardinal of a the set \(X\)). Algorithm 2 has no steps, no comments or results listed (I believe it was not implemented yet). There are no comments about the complexity of algorithms; the step ''\textit{Find a set of cyclotomic cosets \(I\)}'' can be very hard to be accomplished, and the authors propose to run the Algorithm 1 with several important restrictions. Therefore it is not clear how much time was necessary to obtain the tables of codes from Section 5.