Modified Diagram Algorithm
Original Publication Date: 1981-Oct-01
Included in the Prior Art Database: 2005-Feb-12
This article describes an algorithm used to reduce the storage requirements for text or to make a communication channel appear as if it is operating at approximately twice the baud rate. The algorithm encodes each character with a variable length Huffman code based on the previous character in the text stream, or in the case of the trigram mode, the previous two characters. These codes are stored in a decoding/encoding table. Fig. 1 shows a typical coding table called a "state" table because the algorithm is in a particular state based upon the previous one or two characters. There is a state table for every "Previous Character" used to encode the "Next Character". For trigrams there are state tables based upon the previous two characters.