Generalized subspace subcodes with application in cryptology

Most of the codes that have an algebraic decoding algorithm are derived from the Reed Solomon codes. They are obtained by taking equivalent codes, for example the generalized Reed Solomon codes, or by using the so-called subfield subcode method, which leads to Alternant codes and Goppa codes over the underlying prime field, or over some intermediate subfield. The main advantages of these constructions is to preserve both the minimum distance and the decoding algorithm of the underlying Reed Solomon code. In this paper, we propose a generalization of the subfield subcode construction by introducing the notion of subspace subcodes and a generalization of the equivalence of codes which leads to the notion of generalized subspace subcodes. When the dimension of the selected subspaces is equal to one, we show that our approach gives exactly the family of the codes obtained by equivalence and subfield subcode technique. However, our approach highlights the links between the subfield subcode of a code defined over an extension field and the operation of puncturing the q-ary image of this code. When the dimension of the subspaces is greater than one, we obtain codes whose alphabet is no longer a finite field, but a set of r-uples. We explain why these codes are practically as efficient for applications as the codes defined on an extension of degree r. In addition, they make it possible to obtain decodable codes over a large alphabet having parameters previously inaccessible. As an application, we give some examples that can be used in public key cryptosystems such as McEliece.