File Structures, Program Structures, and Attributed Grammars

A language for defining sequential file structures, characterized as nested sequences of records having in common certain keys and types, is presented. "Input schemata" are defined as program skeletons that contain all the necessary control structure to process a specified file. A method for obtaining an input schema from the corresponding file structure definition is given. The method is based on attributed grammars, and has been implemented in the programming language PROLOG. This constitutes a formalization of some aspects of the data-directed program design method of Jackson and Warnier. Examples of applications of this method to business data processing problems such as file updating and report generation are given.