Compacting homogeneous text for minimizing storage space

Abstract

A numerical coding method for saving the storage space for text information is described. Text may be coded at multiple levels—such as at the levels of characters, words, fields, or groups of fields—by repeatedly applying the same encoding procedure. At each subsequent level, the redundantly used common messages found at that level are replaced by a numeric code requiring less storage space. The efficiency of the method depends upon the homogeneity of the text used. The encoding and decoding procedures are described. The storage of the common messages is so arranged that the address of the message can be directly computed from the code, thus simplifying the decoding process. Some empirical data were collected, and significant savings in storage requirements were observed. The analysis of storage requirements is discussed.