Abstract

Title: Towards Value Based Requirements Traceability
Published: September 2005
Authors: Grant Zemont
Abstract: Over the past decade, several studies have been performed that show the importance of requirements traceability (RT) for supporting critical system development activities. While much attention has been focused on how RT should be performed from a technical and process-oriented perspective, little effort has been made to explore the financial benefits and trade-offs of implementing RT as an integral and value-enhancing component of a software development project. In fact, several studies have shown that many IT organizations do not practice RT and therefore must perform exhaustive searches through requirements, design, and code artifacts in order to understand the impact of a proposed change. Other organizations implement basic traceability practices, perceiving them as necessary overhead, but failing to recognize RT as a value enhancing activity. Only a few organizations have a more mature view of RT and understand its value for validating requirements or supporting impact analysis during change management.

Using traceability on projects and using it effectively are two different things. Since 1994, the Standish Group Chaos Report, which as an annual survey of the successes and failures on IT projects, has identified requirements-related problems as several of the leading causes of project failure. While recognizing the importance of the requirements process, many organizations fail to effectively implement RT practices, therefore suggesting that there is a major disconnect between proper or effective utilization of RT and realizing and capitalizing the benefits that it can provide. Traceability affects all the requirements of a system and the subsequent design, code, and tests that are derived from them. It affects all stakeholders in the RT process such as customers, management, developers, and testers. Therefore, RT cannot be simply viewed as an activity to meet a government standard or as an ancillary process to the software engineering domain. As requirement-related practices are crucial to project success and RT supports the requirements process, it is important to understand the actual value that RT can provide both to an individual project and to an organization as a whole, so that informed and business oriented decisions can be made.

This research introduces a framework for assessing the value that RT can provide to an organization in order to support decisions related to the implementation of RT. The Value Based Software Engineering (VBSE) framework, initially proposed and developed by Barry Boehm, is applied to the RT problem. By using the seven facets of this framework in combination with a blend of manual and dynamic traceability techniques, it is posited that an organization can determine how to exploit the implementation of RT practices to maximize both project and organizational value in order to increase the chance of project success. The managed approach that is proposed through this research is called Value-Based Requirements Traceability (VBRT).
Full Paper: [pdf]