McCabe's cyclomatic complexity index is a summary index of binary decisions. It does not distinguish different kinds of control flow complexity such as loops versus IF-THEN-ELSES or cases versus IF-THEN-ELSES. Researchers of software metrics also studied the association of individual syntactic constructs with defect level. For instance, Shen and associates (1985) discovered that the number of unique operands ( n 2 ) was useful in identifying the modules most likely to contain errors for the three software products they studied. Binder and Poore (1990) empirically supported the concept of local software quality metrics whose formulation is based on software syntactic attributes. Such local metrics may be specific to the products under study or the development teams or environments. However, as long as an empirical association with software quality is established, those metrics could provide useful clues for improvement actions. In selecting such metrics for study, consideration must be given to the question of whether the metric could be acted on.
In studying the quality and syntactic indicators among a sample of twenty modules of a COBOL compiler product, Lo (1992) found that field defects at the module level can be estimated through the following equations:
While both equations provide satisfactory results, the findings mean nothing in terms of planning actions for improvement. In a second attempt, which included all 66 modules of the product, Lo examined other syntactic constructs and found the following relationship:
In the model, all three metrics are statistically significant, with DO WHILE having the most effect. The DO WHILE metric included both DO WHILE . . . END and DO WHILE TO . . . . ? Although the R 2 of the model decreased, the findings provide useful clues for improvement. Although it is difficult to avoid the use of IF THEN or to change the number of unique operands, it is feasible to reduce the use of a complex construct such as the DO WHILE or SELECT statement. Upon brainstorming with the development team, Lo found that most developers were having difficulty mastering the DO WHILE construct. As a result, minimizing the use of DO WHILE was one of the actions the team took to reduce defects in the compiler product.
What Is Software Quality?
Software Development Process Models
Fundamentals of Measurement Theory
Software Quality Metrics Overview
Applying the Seven Basic Quality Tools in Software Development
Defect Removal Effectiveness
The Rayleigh Model
Exponential Distribution and Reliability Growth Models
Quality Management Models
In-Process Metrics for Software Testing
Complexity Metrics and Models
Metrics and Lessons Learned for Object-Oriented Projects
Measuring and Analyzing Customer Satisfaction
Conducting In-Process Quality Assessments
Conducting Software Project Assessments
Dos and Donts of Software Process Improvement
Using Function Point Metrics to Measure Software Process Improvements
A Project Assessment Questionnaire