Any of many methods in information theory whereby data can be
coded or receded in order to take advantage of redundancy in that data. In
ordinary alphabetic text the data size can be reduced by a factor of from 2 to
nearly 5 by the use of such methods.