A NEW PERSPECTIVE ON REQUIREMENTS ELICITATION METHODS

Journal of Information Technology Theory and Application, 2003 by Tuunanen, Tuure

ABSTRACT

We open a new area of discussion in requirements elicitation and we argue that a new category of end-users is emerging: wide-audience end-users (WAEU). We propose that current literature does not acknowledge the elicitation needs of WAEUs. To support our argument, we develop a framework consisting of two dimensions, reach and communication, in order to review the literature. Using this framework, we form a perspective on the state of research in the area, and determine that academia is moving in the right direction in the development of new elicitation methods. However, we argue that at the moment, no one method or technique presents an integrated solution. To fill this gap, we present three studies that enable us to argue that WAEUs are reachable, but that more work is needed in the field. We propose some suggestions on how this could be done, and identify three areas needing further research: selection of the participants, representation of requirements across stakeholders, and integration of risk in selecting the elicitation method.

INTRODUCTION

Should not an information system meet the needs of those it aims to serve? To answer this question, researchers have struggled since the dawn of information systems (IS). A separate discipline has emerged from IS to answer these questions: requirements engineering (RE). We want to bring forward a new area of research for the RE community: wide audience end-users (WAEU). We argue that the community is facing a new type of end-users. Wide audience information systems are emerging, such as embedded Java applications for 2.5G and 3G mobile phones or digital TVs for consumer markets. These systems are developed for end-users who are not within organizational reach.

The traditional methods of requirements determination (Davis 1982) no longer assist software engineers in approaching the new type of end-users. To understand what WAUEs needs for IS are we need to include them in the planning and development process (Peffers, Gengler and Tuunanen 2003). Otherwise we will once again face the problems all too familiar with innovation: Prototypes within development process show great promise, but when introduced to markets the products are easily rejected (Peffers and Tuunanen 2002).

RE literature argues that one of the key reasons behind unsuccessful IS projects is a failure in RE (Lyytinen and Hirschheim 1987). Discussion of RE has been lively, focusing on issues surrounding problems in eliciting and managing changing requirements. Requirements are generally specified as something that the product must do or as an objective that it should reach with respect to quality (Robertson and Robertson 2002). If considered in terms of IS, requirements can also be defined as descriptions of 1) how the system should behave, 2) application domain information, 3) constraints on the system's operation, or 4) specifications of a system property or attribute (Kotonya and Sommerville 2002).

Unfortunately, current RE literature does not acknowledge the elicitation method needs of WAUEs. We will argue that the discipline lacks the means to reach WAUEs and to understand their needs. We distinguish through a review of literature two dimensions for filling the gap: reach and communication. Using these two dimensions, we recognize three areas in the RE literature where researchers have tried to push the limits of the discipline and reach WAUE needs. Focusing on these two dimensions, we also want to emphasize the elicitation of requirements versus the determination aspect of previous studies in IS literature.

We use three studies to form a perspective on the state-of-the-research in the area and examine how they approached the issue of understanding and reaching WAEUs. As a contribution of the work, we propose a new perspective on the requirements elicitation methods discussion: eliciting the requirements of wide audience end-users. With the perspective, we attempt to identify the future needs of research if it is to successfully elicit the needs of WAUEs.

We limit our research to a discussion of the dimensions mentioned at a high level of abstraction. We do not seek to establish a way to select distinct methods or techniques for a specific type of IS project. Therefore, we exclude the risk (Mathiassen and Munkmadsen 1986) involved in an IS project, even though we agree that it is an important factor, as Davis (1982) has already argued.

The structure of the paper is as follows. We begin by reviewing the elicitation methods literature, and follow how the discipline has evolved along with progress in the discussion of information systems development. Then we confront the requirements elicitation needs of wide audience end-users, and examine three studies through a framework based on the previous paragraph. After the review, we discuss the results of our study, and present conclusions and future research topics.

REVIEW OF REQUIREMENTS ELICITATION METHODS

Requirements engineering can, probably best, be described as a process. One good description is the overview by Pohl (1994): the three dimensions of requirements engineering. His framework assumes that requirement specifications are developed through a process that leads from vague ideas presented in textual languages and without consideration of agreed upon viewpoints into a desired end state, where there is a common agreement on a set of relatively formalized requirements that serve as a blueprint for IS design and implementation.