Defect Prevention Based on 5 Dimensions of Defect Origin

Abstract

" Discovering the unexpected is more important than confirming the known [7]. In software development, the " unexpected " one relates to defects. These defects when unattended would cause failure to the product and risk to the users. The increasing dependency of society on software and the crucial consequences that a failure can cause requires the need to find out the defects at the origin itself. Based on the lessons learnt from the earlier set of projects, a defect framework highlighting the 5 Dimensions (Ds) of defect origin is proposed in this work. The defect framework is based on analyzing the defects that had emerged from various stages of software development like Requirements, Design, Coding, Testing and Timeline (defects due to lack of time during development). This study is not limited to just identifying the origin of defects at various phases of software development but also finds out the reasons for such defects, and defect preventive (DP) measures are proposed for each type of defect. This work can help practitioners choose effective defect avoidance measures. In addition to arriving at defect framework, this work also proposes a defect injection metric based on severity of the defect rather than just defect count, which gives the number of adjusted defects produced by a project at various phases. The defect injection metric value, once calculated, serves as a yardstick to make a comparison in the improvements made in the software process development between similar set of projects.