辞書配列を利用した非モード方式によるシフトJIS文書圧縮

Abstract

This paper proposes a new data compression method for a Japanese-text file, where the text is written in shift-JIS (JIS X 0208) codes. In the first pass, a dictionary array is built up by the higher frequency both single and multi-byte characters. Then in the second pass, the dictionary items substitute all the registered characters. The code 0xFF is put into a compressed file in front of nonregistered character so as to distinguish non-registered characters from registered ones. It takes O (1) time on a hashing basis to confirm whether each input character is in the dictionary, and to transfer its code to a dictionary item. Furthermore, the run-length coding apply to a sequence of successive spaces for the purpose of accomplishment of the much higher compression ratio. The code 0xFE is used in this coding. A feature of the method is to be a non-modal type of compression.