Skip to Main Content
The devastating effect of errors in adaptive data compression is a long-standing open problem. In this paper LZ'77 is changed theoretically and experimentally observed, such that in a significant proportion of LZ'77 phrases, there is more than one copy of the longest prefix in the compressed file. Once the redundant bits of LZ'77 have been identified, it is exploited for channel coding. For error correction and detection RS (255,255-2e) Reed-Solomon codes are used.