Predicting source-code complexity at the design stage

It is shown how metrics can be used to gauge the quality of source code by evaluating its design specifications before coding, thus shortening the development life cycle. The predictive abilities of several types of metrics, all quantitative, are evaluated, and statistics that can be used to adapt them to a particular project are provided. The study data consist of programs written by students, but, because the programs are large, the professor did not provide the specifications, and the test data were not provided, it is believed that the programs are sufficiently realistic for the approach presented to be generally valid.<>