Croll, Paul R. (1997, American Society for Quality) Computer Sciences Corporation
The Taxonomy-Based Risk Identification approach has been adapted by Computer Sciences Corporation (CSC) for its software process improvement efforts. The initial CSC application of the Risk Taxonomy and its Taxonomy-Based Questionnaire (TBQ) produced disappointing results. The approach then was improved by: (1) focusing on risks having the highest of high (HH) probabilities of occurrence; and (2) paying attention to the link between a process-based approach and promises made to customers about products. The resulting methodology had three phases. Phase 1 was preparation, and it contained eight steps. These included: identification and training of a risk assessment team; identification of project information; completion of a TBQ; and generation of interview questions. Phase 2 was the on-site assessment, a four-step process that included: briefings; brainstorming by project personnel on project products, promises, and risks; interviews; and consensus building by the team on its findings. Phase 3 covered formal reporting and tracking of lessons learned. The methodology was applied to a C/UNIX/Motif Command and Control software development project. The risk assessment team validated six of eight HH risks listed by project members. The team also identified six other HH risks. Although all the risks were known to the project members, the risk assessment exercise helped increase the critical awareness of these issues. In a second application, subject experts and a benchmarking approach were used to identify 23 HH risks in a C/UNIX/Motif software development project. These included 17 of 24 HH risks noted by the project team.
Case study,Software quality,Product development,Process improvement