Research on IB in RE
Although the importance of IB in RE seems clear, there has been little research specifically addressing the information behaviors of those involved with RE, whether Requirements Engineers or other project stakeholders. A search of the Library and Information Science Abstracts (LISA) database found no sources which matched the terms “information behavior” and “requirements engineering” and 28 documents which matched the term “requirements engineering,” suggesting that, although researchers may publish work on RE in LIS journals, they do not explicitly associate that research with IB. A search of the journal Requirements Engineering returned no sources for the term “information behavior,” two sources each for the terms “information seeking” and “information use,” and four sources for the term “information needs.”
Alvarez's (2001) study of the ways in which people describe information systems to be replaced by new systems provides an example of a social constructionist view of IB in RE. She explains that “IS development and requirements analysis interviews are events through which social reality is constructed, more importantly, they are power laden processes” (p. 401). Alvarez's focus on the discursive practices within RE and the social construction of reality echo Tuominen, Talja, & Savolainen's (2005) notion of social constructionism as the ways in which “information practices, actors, and technologies are constructed in discourse and conversation” (p. 331). Writing almost a decade earlier, Gougen (1992) posits a similar understanding of the nature of IB in RE:
However, requirements are emergent, in the sense that they do not already exist in the minds of clients or requirements engineers (or anywhere else); instead, they gradually emerge from interactions between requirements engineers and the client organisation. Moreover, requirements are open: they are always subject to change, because organisations and their contexts continually change. Requirements are also local, in that requirements documents must be interpreted in the context of a particular organisation at a particular time…. Moreover, they are contingent, because they are an evolving outcome of an on-going process that builds on prior interactions and documents. (p. 11, emphasis in the original)
Siddiqi (1994) provides an additional example of a social constructionist view on IB in RE. He comments that “assumptions about things like organizations and society invariably become embedded in the requirements method as it is developed,” going so far as to assert that “reality is socially constructed as a result of interactions among participants in the requirements process” (p. 19).
Poltrock, Grudin, Dumais, Fidel, Bruce, & Petjersen (2003) also specifically discuss IB in RE, saying, “This information space is not a coherent collection readily available to the designers; finding and sharing the information needed to define their product is part of a design team's work” (p. 239). The authors go on to explain the different ways in which information retrieval activities in the project took place individually, collaboratively, using a repository, and via interaction with other people (p. 241). Desouza (2003) also describes interpersonal and computer-based IB activities related to RE, focusing specifically on reasons why the computer-based information repositories are less likely to be used in the software development process (p. 100). Finally, Walsh & Shneider (2002) highlight the importance and danger of the uses of RE practices and the resulting artifacts, saying, “Relying on complete requirement analysis may actually contribute to failure because of overconfidence and because of ignoring risks” (section 4, ¶ 1).
Research related to the topic of IB in RE
Although the preceding section shows that authors report very little research specifically on IB in RE, the many perspectives of IB theory can be found, albeit often unstated, in research on RE and information systems development more generally. Research in RE typically takes one of two approaches. In one approach, researchers investigate methodologies and tools to improve the current state of RE practice, often adopting a cognitivist approach to capturing a finite set of existing, agreed-to requirements (Gemino & Wand, 2004, and Sutcliffe, 1996). The other common approach proposes a more social constructionist view of information systems development in which the requirements are emergent from the social context and various interactions within which they are formed (Galal & Paul, 1999, Goguen, 1992, Rose, 2002, and Truex, 1999). This split mirrors differences among IB theorists who take a philosophical realist perspective on IB (e.g., Belkin, 2005, Kuhlthau, 1991, and Taylor, 1968) and those who view information as socially constructed and contextual (e.g., Agre, 1995, Frohmann, 2004, Hjørland, 2007, Talja, Tuominen, & Savoleinen, 2005, and Wilson, 1997).
In the case of Urquhart's 2001 paper, RE practices are cited as potentially useful in IB work, even though she states in the opening sentence that “any relationship between information behaviour analysis and the structured approach to information requirements analysis of conventional systems analysis seems remote” (Section 1). Through her analysis, Urquhart shows that there are important “similarities in the approaches, despite the differing provenance (software engineering and information science)” (Section 9, ¶ 2). Also, even when the literatures of IB and RE may not specifically refer to each other, at times they use strikingly similar terms and concepts. For example, Davidson states that RE “is characterized by ongoing sensemaking among stakeholders” (2002, p. 329), utilizing a term familiar in LIS. Pettigrew, Fidel, and Bruce (2001) offer an important perspective and caveat on the lack of collaborative or cross-discipline work on IB in RE:
Most researchers who study information behavior are not personally interested in the design of systems and services, and they report their studies to the benefit of other researchers in information behavior. This separation between the “human” side and the “system” side of information behavior is not useful if we believe that information systems and services should be designed to support information behavior and that the design of such systems be based on our understanding of this behavior. (p. 64)
The IB concept of communities of practice also applies to RE research and practice. Davenport & Hall (2002) say that a community of practice “denotes the level of the social world at which a particular practice is common and coordinated, at which generic understandings are created and shared, and negotiation is conducted” (p. 172). While this definition reflects the focus and goals of a typical information systems development team, the authors are careful to explain how the two differ:
Communities of practice are not goal driven (unlike teams and projects), nor are they necessarily deadline driven. Freedom from such constraints makes them, in some circumstances, environments that are more hospitable to sharing and synergy than conventional competing organizational subgroups. (p. 182)
Since project teams do not constitute such communities, of what use is the concept to RE work?
Communities of practice are composed by those people who work as Requirements Engineers, both within organizations and outside organizational boundaries. While it is easy to see how RE work involves information related to a current systems development project, IB related to the meta-information about RE work is less explicitly discussed. The information behaviors of Requirements Engineers reflect many reported in the literature about communities of practice, including Irick's (2007) focus on tacit knowledge within a community, Davenport & Hall's concept of a “corpus of cumulated experience which becomes, in itself, a key artifact in community activity” (2002, p. 176), Savolainen's “institutional activity that consists of more or less formal sets of rules concerning, among other things, what should be considered ‘;proper’ information seeking” (2007, p. 125), and Wenger & Snyder's assertion that as communities of practice “generate knowledge, they reinforce and renew themselves” (2000, p. 143).
While these concepts are of great importance to practicing Requirements Engineers, they receive little attention from researchers. As Cheng and Atlee (2007) report, only 10-15% of research in RE is “evaluation-based” (“whose mission is to assess the state of the practice and evaluate proposed advances to the state of the art”), while the remaining research is “solution-based” (“which emphasizes technological advances that make progress towards solving RE problems; such research is often accompanied by proofs-of-concepts or pilot studies that show the potential of the proposed ideas”) (p. 290).
The juxtaposition of current research addressing IB in RE with the current practice of RE suggests that Requirements Engineers employ theoretical frameworks and software tools which seek to meet others' information needs the wrong way. As Poltrock, Grudin, Dumais, Fidel, Bruce, & Petjersen (2003) and Desouza (2003) note, software developers often prefer to seek information from people, not document repositories. The developers prefer the additional context provided during an interpersonal discussion of a requirement, and current RE software tools do not easily support the documentation of a requirement's contextual setting, meaning, or importance. The RE community of practice has built an “institutional circuitry” (Agre, 1995, ¶ 11) which constrains the possible creation, collection and communication of requirements information, and that circuitry does not appear to meet the needs of other key communities. The focus of RE practice and the functionality of RE tools demand greater attention in order to propose useful, not easy, solutions to these information problems.
Analysis of the current state of research
The lack of research specifically addressing IB in RE would seem to suggest that such research is not of interest, of use, or well-received by researchers or practitioners in LIS or software development. As evidenced by Cheng and Atlee's (2007) review, the majority of research published in RE is focused on incremental improvements to established practices, not paradigm-shifting ideas. RE is still a relatively young field, having formed from the only slightly older field of software engineering. As such, it may be that the RE field is too young to be ready to shift – it may still be trying to establish an initial paradigm. Akkermans & Gordij (2006) question whether existing research merits the description “scientific,” given the manner in which it is currently undertaken.
The popularity of instructional, rather than self-reflective, monographs on RE practice further suggests that RE practitioners are interested in reading about what to do, not about principles underlying what they do. Then again, many Requirements Engineers may simply not recognize the value of research and research findings to their day-to-day work. As one RE messageboard respondent stated when commenting on academic information sources, “The majority of what I've seen is: 1) steeped in academic-speak to the point they are essentially unreadable, 2) aimed specifically at other academics and not real-world practitioners, and 3) don't use a lot of examples that I could apply to work I do” (Seilevel, 2007b). Additionally, the work required to achieve an understanding of IB concepts is likely to be more than most Requirements Engineers are willing to undertake, especially if “they perceive that their world is working without it” (Pettigrew, Fidel, & Bruce, 2001, p. 55).
Although research on IB in RE may be sporadic, research on IB certainly is happening within the LIS community, and research on RE is happening within the software engineering community. Even though the two research domains contain similar terminology and provide fertile ground for inquiry, very little crossover is taking place (see Davidson, 2002). Such isolation of research activities constitutes missed opportunities for IB research, RE research, and RE practice. RE's specific focus on the creation, collection, dissemination and management of information within a social context provides a number of opportunities for the investigation of IB theories, from the more cognitivist to the postmodern. IB's wide range of theoretical viewpoints provide RE research with multiple hermeneutics for the investigation of RE practice and theory. Finally, the principles of IB theory offer RE practitioners valuable insights into their own patterns of behavior, their working contexts, and the mangle of practice in which their work creates and is shaped by information and IB (Pickering, 1995).