Quality measures in design

Abstract

Structure Charts are the third major deliverable of a structured design, following the logical model and the physical model. Code is usually written immediately after the Structure Charts are approved. Many metrics are available for measuring code complexity and some forms of data coupling. Applying similar metrics to the design might be used to find problems in complexity and coupling before interfaces are formalized and code is written. This paper reviews complexity and coupling code measures which could be applied to structure charts. Since there has been little analysis and data collected in this area, it is proposed that relative figures of merit can be derived from structure charts. It is also proposed that in large systems, structure charts with out of normal figures of merit are candidates for detailed review. The general outline of an Automated Auditor is also defined as a tool to assist designers and reviewers in design analysis.