Recovery from dropout errors in LZ77 compressed text

Summary form only given. A dropout in LZ77 compressed text occurs when a portion of the compressed bitstream has been lost. As an example, a bad sector on a hard disk could create a dropout in a stored compressed file. The file may be properly decompressed up to the start of the dropout but then the decompressor can only place variables where the text has been lost. The problem of recovering the variables is a classical one in cryptanalysis which can be attacked in a variety of ways. One technique is to set the variables by matching partial words against dictionary entries. Each word that matches the partial word template provides possible solutions for the embedded variables. To facilitate looking up words which are partially obscured by variables, a hash table is built which contains entries for words containing as many as three variables. Each partial word is hashed and a pointer to the dictionary entry of the full word is then placed into the hash table. The strategy to recover the variables is simply to pass through the data searching for partial words with three or fewer variables. The partial words are matched against dictionary entries via the hash table and a record is kept of the possible solutions for the variables. After the entire text has been examined, those variables with a solution that scores substantially above its competitors are set. Multiple passes are then made through the data in an attempt to recover more variables after their neighbors have been set