Seminar: Defect Management and Prediction

Defects are one major cost driver in the development of software-based systems. A known fact in this area is that the later a defect is found in the development process, the more expensive it is to correct. In the worst case, failures of the system during use caused by defects may trigger huge costs and may result in human casualties.

Defect management and prediction is a common best practice in the area of engineering systems and software to allow controlling defects, to get transparency regarding where defects occur most often, to identify root causes of defects, and, finally, to predict the number of defects.

This seminar gives a basic introduction to defect management and prediction. It presents common defect classification schemes and introduces two concrete approaches for finding defects earlier in the development process and predicting defect contents: Defect Flow Models support an organization in analyzing how many defects of which type have slipped through the quality assurance stages of its development process. HyDEEP is an approach for defect risk management and supports organizations in predicting defects at the planning stage of a development project based on historical data and expert knowledge.

Introduction to Defect Management

• Motivation and Impact
• Consequences of Software Failures
• Effective Quality Management
• Defect Classification
• Defect Flow Models
• Defect Prediction

Defect Classification

• Terminology
• Structures for Classification
• IEEE Standard Scheme
• Hewlett-Packard (HP) Scheme
• Orthogonal Defect Classification (ODC)

Defect-Flow Modeling

• Basic Defect Flow Models
• Usage Scenarios for Extended Models
• Implementing Defect Flow Models
• Industrial Application Examples

Exercise: Creating a Defect Flow Model

Defect Prediction

• Basic Approaches
• HyDEEP Approach
• Application Scenarios
• Industrial Studies
• HyDEEP Tooling

Summary and Conclusions

• Benefits versus Cost
• Lessons Learned