One task that comes up occasionally in programming, particular in financial or text-to-speech applications, is spelling out a numeric value is words. While the algorithm for doing this in English is relatively simple, it cannot be generalized easily for use with other languages. In fact, the ways in which numbers are spelled out in words vary so much from language to language that the only practical and general approach is a rule-based approach.

This paper describes a simple language for describing number formatting procedures that not only solves the international number-spellout problem, but also can be used for a variety of other number-formatting tasks too complicated to be handled by standard number formatters. Examples include formatting numbers using numeration systems other than Arabic numerals, formatting ordinal numbers, formatting values in major and minor units (e.g., a number of seconds as hours, minutes, and seconds), formatting numbers with fractions (e.g., NYSE stock prices), and formatting numbers with unit designations that change with the number's magnitude (e.g., 1 as "1 m", 1,000 as "1 km", and 0.001 as "1 mm").