Abstract: A method is disclosed for computer composition of Nastaliq script and similar Ruka'ah script of the Urdu group of languages (Arabic, Persian, Punjabi, Pushto, Sindhi) in full conformance to rules of a script. Basic constituent elements of the characters of scripture are stored in a computer memory and these elements in response to user command are joined with each other, as required, to compose scripture according to a set of rules defined and stored in a rules dictionary. Combinations of characters that do not conform to the general rules contained in the dictionary are treated and executed as exceptional cases by the computer program. Moreover, an efficient mapping of Urdu characters to the existing QWERTY keyboard is achieved by matching character frequency of occurrence to finger agility, by pattern recognition features of similar characters, and by rationalizing the overall finger workload.