Skip to main content

ODC Chapter: Handbook of Software Reliability Engineering

Ram Chillarege
Michael Lyu, Editor in Chief, IEEE Computer Society Press, ISBN 0-07-039400-8, pp. 259-400, 1996

info

This excellent book was published with 17 chapters, covering the area of Software Reliability Engineering. Ranging from modeling, operational profiles, prediction, testing, simulation, fault-trees, neural nets, Ram wrote Chapter 9, on Orthogonal Defect Classification. The pdf of the chapter is available on this site.

PDF​

PDF: ODC Chapter SRE Handbook, IEEE Computer Society Press

Chapter 9​

Orthogonal Defect ClassificationPages Pages: 359-400.

Sections:​

  • 9.1 Introduction
  • 9.2 Measurement and Software
    • 9.2.1 Software defects
    • 9.2.2 The spectrum of defect analysis
  • 9.3 Princples of ODC
    • 9.3.1 The intuition
    • 9.3.2 The design of orthogonal defect classification
    • 9.3.3 Necessary condition
    • 9.3.4 Sufficient conditions
  • 9.4 The Defect-Type Attribute
  • 9.5 Relative Risk Assessment Using Defect Types
    • 9.5.1 Subjective aspects of growth curves
    • 9.5.2 Combining ODC and growth modeling
  • 9.6 The Defect Trigger Attribute
    • 9.6.1 The Trigger Concept
    • 9.6.3 Review and inspection triggers
    • 9.6.4 Function test triggers
    • 9.6.5 The Use of Triggers
  • 9.7 Multidimensional Analysis
  • 9.8 Deploying ODC
  • 9.9 Summary
  • Problems

Introduction​

One of the perpectual challenges for measurement in the software development process is to provide fast and effective feedback to developers. Traditional techniques, although they provide good report cards, seldom have the fine insight into the process or product to truly guide decision making. Thus it is not uncommon to witness decisions guided more by intuition than by true measurement, analysis, and engineering.

Orthogonal Defect Classification (ODC) makes a fundamental improvement in the level of technology available to assist software engineering decisions via measurements and analysis. This is achieved by exploiting software defects that occur all the way through development and fiel use, capturing a significant amount of information. ODC extracts the information from defects cleverly, converting what is semantically very rich into a few vital measurements providinga firm footing upon which sophisticated analysis and methodologies are developed. Thus it is possible to emulate some of the detail and expansiveness of qualitative analysis with the low cost and mathematical tractability of quantitative methods.