Towards the ultimate text compression method for help systems

State-of-the-art in the area of text compression for practical applications seems to be rapidly approaching its theoretical limit. This limit was estimated at 1.2 bits per character [6] (on the basis of the ability of a human to predict textual data input), and it establishes the ultimate goal of every investigation. This paper reports a research which led to an algorithm displaying a good-faith agreement between the objective of compression and the performance of contemporary computational resources, bringing the efficiency as high as 1.6 bits per character. It presents a step-by-step reasoning which directed the study toward the discovery of an intuitively perfect technique and discusses various heuristic rules that have been utilized to amplify the rudimentary engine.