We view the creative process in RE as having both emergent and design characterisations.
First, RE has been described as an ill-structured problem solving process. The ill-structuredness of the requirements problem can be characterised through the open-endedness of the problem, poorly understood problem context, existence of multiple domains, complexity and dynamics of social interactions, organisation structures, business processes and technologies involved (Guindon, 1990b; Conklin, 2005; Nguyen and Swatman, 2006). In ill-structured problem solving, the understanding (and discovery) of the problem and structuring of the solution are intertwined. The problem solver (i.e. the systems analyst) moves between different problem areas in search of a possible solution, interacts and communicates his or her understanding with other stakeholders, and responds to the emergent situation. Therefore, both the problem space and the solution space progressively evolve as the systems analyst gains more knowledge and responds to the stimuli produced by the social, business and technical complexity and dynamics. This is similar to a description of the creative design process: ‘The designer operates within a context which partially depends on the designer’s perceptions of purposes, constraints and related contexts. These perceptions change as the designer explores the emerging relationships between putative designs and the context and as the designer learns more about possible designs’ (Gero, 1996).
Second, RE should be seen as an art involving two different acts — articulating and documenting user requirements (i.e. describing the real world situation) as well as designing new business practice (enabled by a new system) by suggesting changes to the current situation. These acts are referred to, respectively, as analysing and modelling As-Is and To-Be requirements. However, the literature tends to focus on the former more extensively. Requirements elicitation, modelling and communication are fundamental activities (Nuseibeh and Easterbrook, 2000; Loucopoulos and Karakostas, 1995; Pohl, 1994). The first activity focuses on the acquisition and articulation of the user requirements. The second activity focuses on the representation and documentation of the requirements in various formats and perspectives. The third activity aims at requirements communication, negotiation and validating a correspondence between the requirements specification and the real world problem. We acknowledge that these fundamental systematic activities are still required in the development of To-Be requirements but we advocate that creative thinking plays a crucial role in envisaging and designing future information and communication technology (ICT) enabled business practice, especially if we aim at innovating in the business and creating new, significant added value through it. The invention of To-Be requirements, which is a largely missing activity in the current RE literature, is a key to envisaging and designing ICT enabled innovative business practice (Robertson, 2002, 2005). In addition, Simsion’s (2006) investigation into data modelling in practice characterised data modelling as a creative design process although it is widely perceived to be a descriptive representation process.
Therefore, RE is a creative, emergent design process. In the next sub-section, we look at different views of the RE process and relate them to their counterparts in design studies and creativity research.
Nguyen and Swatman (2003, 2006) distinguished two views of the RE process that are held by members of the RE community:
The first view describes a systematic, structured and evolutionary process. Though detailed descriptions of the RE process may vary, essentially the requirements model is pictured as continually, incrementally structured and refined through a cyclic systematic process (e.g. Alexander, 1998; Loucopoulos and Karakostas, 1995; Kotonya and Sommerville, 1998).
The second view describes an opportunistic, constructivist process consisting of structuring and insight-driven restructuring of the requirements model. These opportunistic cognitive behaviours and insight-driven reconceptualisations of the problem space by the systems analyst are important in handling the emergent problem space and partial solutions (Guindon, 1990a; Visser, 1994; Khushalani et al., 1994; Nguyen et al., 2000; Robillard, 2005).
The existence of these two views of the RE process is reminiscent of the observation made by Dorst and Dijkhuis (1995) that there exist two views of the design process in the design studies community. The first view describes a rational problem solving process characterised by structured search and information processing in the problem space (Newell and Simon, 1972; Simon, 1969). The second view is constructivist and describes the design process as a reflective conversation with the situation (Schön, 1996).
Dorst and Dijkhuis (1995) have argued that the former characterisation describes the problem solving process for structured and fairly clear cut problems whereas the latter describes the conceptual design stage for ill-structured problems. Based on the discussion above, the latter matches the characteristics we attribute to the creative part of the RE process. We further note that these two descriptions of the design process are analogous to the two descriptions of the RE process. We see this analogy as a manifestation of two ‘forces’ in RE: the enforcing of a systematic structured process to avoid a chaotic error-prone process, as opposed to the recognition and taking advantage of opportunistic cognitive behaviours and heuristics of professionals (in dealing with the emergent problem space). Both of these two forces are essential in RE problem solving; a good balance of them is required.
Boden (1991) has described the creative process as an internal process of exploration and transformation of conceptual spaces in an individual mind. However, understanding how this internal process, which actually happens in the individual mind, has long been a challenging topic in creativity research. There exist numerous models to describe the creative process. Shneiderman (2000) described three creative process models: inspirationalist, structuralist and situationalist.
The inspirationalist views the creative process as passing through four phases of preparation, incubation, illumination and verification, as in the model of Wallas (1926), and as unpredictable insight and associated restructuring of the problem space in Gestalt psychology (Mayer, 1992; Ohlsson, 1984). Common creativity enhancing techniques, such as lateral thinking, divergent thinking, six thinking hats, and free association, are often adopted to promote inspirational creativity.
The structuralist views the creative process as a more focused and structured effort to generate and evaluate ideas as in, for example, Osborn-Parnes’ Creative Problem Solving CPS (Daupert, 2002; Osborn, 1979) and the Directed Creativity Cycle (Plsek, 1997). A cyclic process of divergent brainstorming and convergent thinking phases is included in these models to stress a balance between imagination and the analytical aspects of creativity.
The situationalist emphasises the social interactions between individual problem solvers and thus stresses the collaborative nature of the creative process. Three out of the four phases of collecting, relating, creating and disseminating in the creative process model of Shneiderman (2000) are designed to directly support collaboration and communication of information and ideas in the creative process. In this vein, a group of researchers at the University of South Australia extended CSCW (Computer Support Collaborative Work) theories in order to develop ICT-enabled supportive workplace for creative teams (Blackburn et al., 2005).
We have two observations. First, the inspirationalist and the structuralist tend to focus on the creative effort by individuals while the situationalist emphasises the collaboration between them. Second, the inspirationalist tends to focus on how the creative process actually occurs while the structuralist and situationalist tend to focus on how the creative process should be undertaken. Based on a synthesis of creativity models from creativity research and creativity research in RE, Nguyen and Shanks (2006) suggested integrating the different views through a collaborative process consisting of cycles of structured building up and opportunistic restructuring of the requirements model. This process needs to be further developed and refined. Overall, we conclude that the differences between different descriptions of the creative process that exist in the RE, design studies and creativity research communities manifest different styles of creative thinking and cognition. Our conclusion points to the need for, and difficulties in, integrating the different creative process views. It is therefore important to further explore creative thinking and cognition in the RE process and the question arises as to which research method(s) would be most appropriate to pursue this exploration.
To explore creative thinking and cognition, it is important to obtain data about the process and to reconstruct what happens in the mind of systems analyst. As highlighted earlier, a major challenge is that systems analysts can not describe accurately how they developed solutions for problems they faced (Lubars et al., 1993; Hofmann and Lehner, 2001). A similar challenge exists in design studies; designers ‘cannot articulate what kind of expertise they use in designing and how’ (Suwa et al., 1998). Researchers in design studies have been using protocol analysis, an empirical method in cognitive psychology, to examine the design process. In the next section, we describe and discuss the potential use of protocol analysis in exploring the creative process in RE.