Application of Formal Grammars for Encoding Information Sources

To increase the efficiency of data compression methods, we suggest describing the structure of data by means of formal grammars. This approach helps to compress even small files for which known adaptive and nonadaptive codes are usually ineffective. Based on this approach, some data compression systems for data bases and program libraries in several programming languages have been built and experimentally investigated. The codes constructed permit one to increase the compression ratio by 10-30% compared to other methods.