Fault links represent relationships between the types of mistakes made and the type of module being developed or modified. The existence of such fault links can be used to guide code reviews, walkthroughs, allocation of verification and validation resources, testing of new code development, as well as code maintenance. We present an approach for categorizing code faults and code modules, and a means for examining relationships between the two. We successfully applied our approach to two case studies.
Jane Huffman Hayes, Inies C. M. Raphael, Vinod Kum