Introducing Requirements Engineering: User Needs and Requirements as a Basis for Product Development

The overall goal of the thesis is to investigate how organizations can introduce requirements engineering (RE) into product development by applying a systematic process improvement approach. The results are based on four longitudinal case studies that were carried out using an action research method. The thesis provides new knowledge about RE process improvement. It offers a set of basic RE practices that describes where organizations can start improving their RE processes. In addition, the thesis provides a set of process improvement guidelines. These guidelines can serve as a checklist for organizations that want to introduce requirements engineering into their product development. 1. BACKGROUND AND MOTIVATION Developing products that satisfy the needs of its intended users is crucial to successful growth and increased profit in many companies. Discovering real user needs, transforming them into user requirements and using these requirements as a basis for product development work can, however, be a challenging task. For example, the Standish Group’s research indicates that lack of user involvement and incomplete requirements and changing requirements are the major reasons why information technology projects do not deliver all of their planned functionality on schedule and within budget [10]. Requirements engineering (RE), a branch of software engineering offers models, methods, and practices for organizations that want to define and manage requirements systematically. Systematic requirements definition and management have long been recognized to be a crucial part of software systems development. Laura Scharer pointed out already in 1981 that good requirements definition can be an important factor in the success of a system development projects [8]. Similarly, Frederick Brooks states in his classic essay that the most important function that the software builder performs for the client is the iterative extraction and refinement of the product requirements [2]. Requirements engineering became a recognized discipline of its own in the early 1990’s [7]. In theory, RE applies proven principles, techniques, languages, and tools to help analysts understand a problem or describe a potential product’s external behaviour [3]. However, there seems to be little use of the concepts, principles and methods developed by RE researchers. According to Stevens et al., most organizations handle user requirements poorly [11]. Nusebeih and Easterbrook also point out that many delivered systems do not meet their customers’ requirements due, at least partly, to ineffective requirements engineering [4]. Moreover, Sommerville and Sawyer argue that very few organizations have an explicitly-defined and standardized RE process [9]. 2. RESEARCH QUESTIONS The overall goal of this work is to investigate how organizations can introduce requirements engineering into product development by applying a systematic process improvement approach. The fundamental assumption of this thesis is that the quality of a product depends on the characteristics of the processes used for producing the product. Therefore, we selected the process improvement approach to be used as a foundation for the thesis. This thesis is based on the Requirements Engineering Good Practice Guide (REGPG) developed by Sommerville et al. [9]. The REGPG is a framework for incremental RE process improvement, and it contains sixty-six good RE practices covering eight RE areas [9]. We selected the REGPG as a basis for this research because in comparison with other software process improvement (SPI) frameworks such as the CMM, the REGPG is unique in focusing specifically on RE processes. Furthermore, it provides detailed guidance on a set of good RE practices and offers practical information for RE process improvement. Additionally, as far as we know, no empirical studies conducted on the REGPG have been published. From the research point of view, this motivated us to evaluate the REGPG by using it in real product development organizations. The research problem “how organizations can introduce RE into product development” is defined into the following four research questions: • How does the REGPG support organizations in improving their RE processes? • What are the basic RE practices where organizations can start improving their RE processes? • Which factors support the introduction of requirements engineering? • Which challenges can organizations confront when introducing RE to product development? 3. RESEARCH METHOD The need for deep understanding on how organization can introduce requirements engineering into product development calls for a qualitative research approach, and more precisely an action research method. According to Avison et al., a particular strength of qualitative methods is their value in explaining what goes on in organizations, and action research can address complex real-life problems and immediate concerns [1]. In addition, qualitative methods permit the evaluator to study selected issues in depth and detail [5]. Also, Colin Potts suggests an “industry-aslaboratory” research approach where researchers identify problems through close involvement with industrial projects, and create and evaluate solutions in an almost indivisible research activity [6]. According to Potts, this lets researchers emphasize what people actually do or can do in practice, rather than what is possible in principle. The research results are based on four longitudinal case studies (Table 1). All the case organizations were product development units of medium-size or large Finnish companies. These companies focus mainly on market-driven products. Three of the case organizations develop interactive systems, and one of them develops products that have both real-time embedded and interactive components. Table 1. Duration of the research co-operation with the case organizations Organization Period Duration A Feb 1999 – Sep 2002 44 months (3,7 years) B Feb 1999 – Sep 2002 44 months (3,7 years) C Jun 2000Nov 2002 29 months (2,4 years) D Aug 2001 – Sep 2002 14 months (1,2 years) The research results are based on the data collected through observations, informal conversations, formal interviews, and analysis of both RE process documentation and requirements documents. The researcher interviewed project managers, product managers, domain experts, usability experts, product development managers, product development engineers, and persons who were responsible for coordinating RE process improvement. The total number of interviews was 42, eleven of which were group interviews. The total number of different informants was 51. The data was analyzed using the grounded theory research method by Strauss and Corbin [12]. The aim of the grounded theory method is to help researchers build theory through qualitative data analysis. It also offers techniques and procedures to those researchers who want to do qualitative analysis but who do not wish to build theory.