Abstract

[This abstract is based on the authors' abstract.]When the work patterns of software development and maintenance projects are analyzed, a surprising hypothesis emerges. Software quality is so poor that productivity is much lower than it should be. Poor quality shows up in three major software economic problems: 1) canceled projects that are never released due to poor quality; 2) schedule delays due to poor quality extending test duration; and 3) excessive work on finding and fixing bugs, which often exceeds 60 percent of total software effort. The amount of software effort spent on software projects that will be canceled due to excessive error content appears to absorb more than 20 percent of the U.S. software workforce. In addition, about 60 percent of U.S. software engineering work time centers on finding and fixing errors that might have been avoided. Finally, software schedules for major applications are about 25 percent longer than they should be due to poor quality expanding testing intervals.