Codes correcting phased burst erasures
Osnat Keren, Simon N. Litsyn · IEEE Transactions on Information Theory · 1998
We introduce a family of binary array codes of size t/spl times/n, correcting multiple phased burst erasures of size t. The codes achieve maximal correcting capability, i.e., being considered as codes over GF(2/sup t/) they are MDS. The length of the codes is n=/spl Sigma//sub l=1//sup L/(/sub l//sup t/) where L is a constant or is slowly growing in t. The complexity of encoding and decoding is proportional to rnmL where r is the number of correctable erasures, and m is the smallest number such that 2/sup t/=1 modulo m. This compares favorably with the complexity of decoding codes obtained from the shortened Reed-Solomon codes having the same parameters.