Arithmetic coding procedures normally represent the output data string as a binary fraction within the unit interval (0,1). As is explained in "An Introduction to Arithmetic Coding" by Langdon, Jr. in the IBM Journal of Research and Development, Volume 28, No. 2, March 1984, arithmetic coding is related to the process of sub-dividing the unit interval. This sub-division is achieved by marking along the unit interval code points Cn for each symbol within the source alphabet, each code point being equal to the sum of the probabilities of occurrence of the preceding symbols. The width or size An of the sub-interval to the right of each code point represents the probability of occurrence of the source data string up to the corresponding symbol (cf. FIG. 1 of that publication).