Programs to process trees, representing program structures and data structures

Abstract

AbstractUtility programs discussed in this paper were initially developed to automate the ‘drawing’ of the non‐binary trees which occur when using a Jackson type methodology for program design; these trees represent either data structures or program structures, showing the selective, sequential and iterative relationships between nodesFirst a language was developed to allow information about tree structures to be input to the programs. Techniques and algorithms are described which enabled utility programs to be coded to process this data; transforming it into printable diagrammatic representation of trees; pseudo‐code representation of program trees; and PL/I or Cobol programsSecondly a merge operation for trees had to be defined, reflecting the step in Warmer's and Jackson's methodologies of combining data tree structures together to provide a program structure. Further algorithms are described which enable this step to be performed automatically, they include checks that each particular merge can be performedBecause the aims and concepts have been fluid some of the ideas have been discarded, and programs have been rewritten. Valuable insights were gained from persuing both the rejected and the acceptable ideas, these are described in this paper.