Este

Previous screen

1-04-20 Applying Software Quality Tools and Metrics
Larry Runge Payoff
The IS department can better manage software quality problems by identifying and eliminating problem sources, rather than treating the symptoms of poor quality. This is classic Deming manufacturing quality technique applied to the software process. Software quality tools and metrics that can be used to some extent in every software organization are described, along with a simple economic formula for governing a software quality program.

Introduction
According to statistics provided by the STAR ’93 International Conference on Software Quality, 53% of all software applications are delivered with faults that would have been discovered had the application been tested before delivery. This is a significant error ratio in any profession, and when software defect rates are compared to those in the manufacturing arena, it demonstrates that there is substantial room for improvement. Unlike in manufacturing, even defining quality as it applies to software is often difficult. Once defined, the application of the proper metrics to measure quality is nowhere near as precise as in other professions where products are created, fabricated, and then distributed. Another complication is that software can be viewed as both a product and a service. The term quality, as applied to software, is often ambiguous and ill-understood. Software is more closely aligned with intellectual products than it is to precision machining, for example. Software quality initiatives generally fall into five categories, which form a logical progression: no measurements, squeaky wheel metrics, multiple metrics, problem determination, and elimination of sources of problems. This last category, elimination of sources of problems—which is the Deming Cycle transferred from hardware (i.e., manufacturing) to software—is where IS organizations need to be. To get to this level, the IS organization must make effective