Histogram

Figure 5.5 shows two examples of histograms used for software project and quality management. Panel A shows the defect frequency of a product by severity level (from 1 to 4 with 1 being the most severe and 4 the least). Defects with different severity levels differ in their impact on customers. Less severe defects usually have circumventions available and to customers they mean inconvenience. In contrast, high-severity defects may cause system downtime and affect customers' business. Therefore, given the same defect rate (or number of defects), the defect severity histogram tells a lot more about the quality of the software. Panel B shows the frequency of defects during formal machine testing by number of days the defect reports have been opened (1 “7 days, 8 “14, 15 “21, 22 “28, 29 “35, and 36+). It reflects the response time in fixing defects during the formal testing phases; it is also a workload statement. Figure 5.6 shows the customer satisfaction profile of a software product in terms of very satisfied, satisfied, neutral, dissatisfied, and very dissatisfied. Although one can construct various metrics with regard to the categories of satisfaction level, a simple histogram conveys the complete information at a glance.

Figure 5.5. Two Histograms

Figure 5.6. Profile of Customer Satisfaction with a Software Product

As the examples show, the measurement scale of the data is either interval, ratio, or ordinal (reference the level of measurement discussions in section 3.2 of Chapter 3). If the measurement scale is nominal (e.g., types of software and models of development process), the ordering of the X-axis in a histogram no longer has significance. Such charts are commonly referred to as bar charts. Both histograms and bar charts are frequently used in software development.