Symbol ranking text compressors: review and implementation

Abstract

SUMMARY Methods used by Shannon in 1951 when investigating the information content of English text areshown to lead to a class of 'symbol ranking' text compressors. Included in this class are some recent compressors which combine high performance with completely novel techniques. These are described under the aegis of the symbol ranking classification. Some new compressors, deliberately designed as symbol ranking, are then described. One of these has operated at over 1 Mbyte per second in software, and should be suitable for hardware implementation at tens of megabytes per second, or higher.