Classification of defects in health software
|Publication Date:||1 January 2018|
This document identifies a defect classification system, or taxonomy, that can be used to classify the types of defects that might exist in software. The taxonomy applies to defect root causes in all types of software (including third-party software, medical device software, test software, and manufacturing software) throughout the software development lifecycle.
The classification system is meant to be neutral with respect to
• programming language,
• intended use,
• risk (the consequences of failure), and
• regulatory status.
The use of a common taxonomy allows industry-wide aggregation of defect occurrence data that can be used to improve software quality. Some examples of ways in which this may be accomplished are presented as use cases in Annex A.
The way in which the standard is to be used is not prescribed. The taxonomy does not attempt to classify the severity of defects, as the consequences of any defect can only be evaluated within the context of the software's intended use. It does not attempt to describe methodologies for analyzing root cause, managing defect resolution, or assigning risk. It does not attempt to categorize defects in quality system processes.
Several annexes are included for information. Annexes B, C, and D are provided to illustrate how the taxonomy can be used to categorize various types of software problems. Annex E provides examples of cause analysis using the taxonomy. Annex F provides a reference table of all defect categories and codes, and Annex G provides the rationale for the types of defects that are in scope for the standard.