Software development monitoring graphs
Murat M. Tanik · ACM SIGSOFT Software Engineering Notes · 1980
Two different software tools that can be utilized in software development projects by management, systems analysts, and programmers are developed. These tools are collectively called monitoring graphs. The first type of monitoring graph is called a complexity graph. It is used to represent the complexity of a program or a system of programs. The second type of monitoring graph is called a development graph. It is used in the control of a software project.These monitoring graphs basically look like Kiviat graphs which are used in performance evaluation studies. Fundamentally, a complexity graph is a circle with eight axes. The values to be represented on each axis are obtained from a set of prediction equations which are generated through multiple linear regression analysis. These complexity graphs can further be grouped into three different types depending on the type of data represented on each axis. These types are individual complexity graphs, cumulative complexity graphs and average complexity graphs. Structurally speaking, a development graph is the same as a complexity graph with exception of the information represented on each axis. In a development graph, each axis represents quantities such as percentage of time spent on debugging, percentage of time spent on testing etc. Development graphs are grouped into two different types depending on the function they serve. These types are program development graphs and system development graphs.The userulness of the complexity graphs is tested on a set of COBOL and FORTRAN programs. For development graphs, development patterns of 69 COBOL programs written by two skilled professional programmers are traced for 33 weeks producing weekly patterns of the development graphs to observe the usefulness and efficiency of this method.