Software Process Improvement Applied to Requirements Engineering

Abstract: To maintain and increase competitive advantages, software organizations must continually strive to refine and improve their development processes. Software process assessment and improvement is the means by which development organizations assure and improve their processes, practices and tools to support their specific needs.

Several well known software process improvement (SPI) frameworks exist, which can be classified into two main categories: inductive and prescriptive. Inductive methods, such as the Quality Improvement Paradigm (QIP), take their starting point in a thorough understanding of the current situation, basing improvement efforts on the issues most critical to the specific organization. Prescriptive frameworks, such as the Capability Maturity Model Integration (CMMI) and ISO/IEC 15504, take an approach based on a set of best practices that has proven successful in other organizations. The improvements that shall be carried out are then established by comparing the studied organization against this set of practices, not however taking any consideration to that particular organization’s specific needs. Extensive SPI frameworks, such as CMMI and SPICE, are viewed by many as too large to comprehend and implement, demanding very considerable resources and long-term commitment. This makes it difficult for organizations to initiate and perform assessment and improvement efforts, as cost and time are crucial considerations.

This thesis presents an introduction to software process improvement and research that addresses some of the shortcomings of existing SPI frameworks.

A software process improvement proposal, consisting of two levels of use cases, targeting early deliverables as a means of project control, is presented. The proposal is cost effective in that it incurs low additional resources and was found to be easily integrated into the existing process in an example at Volvo 3P. Use cases are also addressed in a quality evaluation at Volvo Car Corporation, where a set of quality criteria based on earlier research is presented and applied. Based on a statistical analysis of the defect data, improvements are proposed to both research on use case writing guidelines and industry practice.

Finally, a framework for software process assessment and improvement planning is presented together with two cases of industrial application at Volvo Technology. The framework contains both assessment and improvement planning capabilities, explicitly developed to be light weight in nature. This and the fact that both activities are carried out on the basis of the organization’s experience and knowledge enable organizations to ensure management commitment to improvement initiatives and help in lowering the initiation threshold for performing process improvement.