A Framework for Understanding Decision Support Systems Evolution David R. Arnott School of Information Management & Systems Monash University Melbourne, Australia Email: david.arnott@sims.monash.edu.au Abstract Terms such as ‘adaptive’ and ‘evolutionary’ capture the organic nature of the development of a decision support system (DSS). However, they are rarely defined in DSS research and their meaning varies widely in the research literature. The aim of this paper is to contribute to decision support systems theory by clarifying the nature of the evolutionary process of a DSS. Using insight from the theory of evolution and prior DSS research, a framework for understanding DSS evolution is developed based on the aetiology, lineage and tempo of evolution. The aetiology of DSS evolution is discussed in terms of cognitive and environmental triggers, an important distinction for managing DSS projects. The lineage of DSS evolution is viewed as occurring within an application and between applications. In terms of the time pattern of evolution, DSS are considered capable of continuous, punctuated, and quantum evolution. The descriptive validity of the framework is demonstrated by using it to classify evolution in a number of published DSS studies. Keywords Evolutive Design, IS Development Methods and Tools, DSS, Research Frameworks. INTRODUCTION Decision support systems (DSS) are computer-based information systems that are designed with the express purpose of improving the process and outcome of decision making. Evolutionary development has been central to the theory of decision support systems since the inception of the field. The functionality of DSS is thought to evolve over a series of development cycles where both the client and the systems analyst are active contributors to the shape, nature and logic of the system. Developers of DSS need to adopt an evolutionary approach because they generally address ill-structured management decisions. It is almost impossible a priori to specify the system requirements in such an environment and the initial versions of the system will help to clarify these requirements. The environment of DSS is subject to significant change and so even if the system requirements have been specified with some accuracy at the start of the project they are likely to change significantly over time. This paper presents research that is best described as a conceptual study (Galliers, 1992). The aim of this paper is to contribute to decision support systems theory by clarifying the nature of the process of evolution within a DSS. This clarification is based on adapting theories from scientific disciplines including biology, palaeontology, geology and physiology. The paper is structured as follows: first, scientific theories of evolution that may be relevant to DSS are discussed. Following this, current theories of DSS evolution are reviewed. Evolutionary and DSS theories are then used to develop a framework for understanding DSS evolution. Case studies of DSS development from the research literature are fitted to the framework to demonstrate its descriptive validity. Finally, some concluding comments are made that may be of value to both researchers and practising systems analysts. EVOLUTIONARY THEORY AND DSS Insight into the nature of DSS evolution can be gained from evolutionary theory in a number of sciences In biology, evolution is “change in the properties of populations of organisms that transcend the lifetime of a single individual” (Futuyma, 1986, p 7). Biologists make a distinction between ontogeny, the changes in an individual over its lifetime, and evolutionary change over generations. Much of what is termed evolution in DSS is clearly ontogeny although cross generational change is also common. DSS are not capable of sexual reproduction but in a sense they do progress through generations where some aspect of the decision logic (which may be thought of as analogous to a genotype) is passed to the next generation. In the language of information systems, a change in generation is a major version change (e.g. version 1.0 to 2.0). Adaptation of individual DSS can be viewed as a minor version change (e.g. version 1.1 to 1.2). The difference between minor and major version changes, or evolution and ontogeny, is a matter of systems analyst and user perception. Darwin (1859/1996) conceptualised evolution as the cumulative result of a very large number of very small changes that occur over a very long period of time. This gradual continuous process is driven by the blind process of natural selection. Biological evolution is not a goal driven process, and an adaptation only becomes successful if the individuals concerned are able to reproduce and pass the feature to later generations who in turn are able pass the feature to their descendants. In Darwin’s words: “.. natural selection acts by life and death, - by the preservation of individuals with any favourable variation, and by the destruction of those with any unfavourable deviation of structure” (Darwin, 1859/1996, p 159). Budd & Coates (1992) found that evolutionary changes in one time period may be reversed in the next and coined the term non-progressive evolution to characterise this phenomenon. Non-progressive evolution is convincing evidence for the lack of any evolutionary goal other than survival. The modern approach to evolution represents a synthesis of many disciplines but still maintains natural selection as an important mechanism. Selection needs variation between individuals to work and it is well accepted that the major source of biological variation is gene mutation. Random genetic drift and geographical isolation are also important factors in variation (Price, 1996). DSS are man-made systems that are created from a design process and therefore any evolution of a DSS is produced by artificial rather than natural selection. Normally, in an artificial selection process the designers have some goal or desirable state in mind. They consciously and deliberately select those features that support or promote the goal between generations. An example of biological artificial selection is the breeding of thoroughbred race horses. Although DSS are the product of a process of artificial selection, they are the class of information system whose development pattern is closest to natural selection. Because of the unstructured nature of managerial judgement, a DSS does not have a goal that is determined to the same degree that an animal breeder has when attempting to breed a derby winner. Further, the goal of the DSS project is likely to change as a result of the development and use of the system. This means that although DSS evolution is to some extent goal driven, the goal itself is subject to evolutionary processes. This system-goal evolution may even be subject to radical change. Many DSS may have no goal other than to support a particular manager. Adaptation can be best viewed as an mechanism of evolution. It is the process of change whereby a feature of the population is improved in relation to some function. In some sciences, for example physiology and psychology, it is an individual rather than a population effect, and can be described as a phenotypic adjustment to a change in the environment. In a similar way DSS adaptation occurs during the life of a system version. Another important evolutionary agent is speciation, the development of one or more new species from a common stock. In taxonomy, the discipline of scientific classification, a species may not be classified in the same manner as in evolutionary biology and may be determined by other factors, the most important of which is morphology. Evolutionary taxonomy allocates organisms into a classification hierarchy of species, genera, families, orders and classes. Alter’s taxonomy of decision support systems (Alter, 1980, chap. 2) classifies DSS applications at the genera level with decision support systems as a type of information systems at the family level. This taxonomy is shown in Table 1. It has been widely used in DSS research and although formulated in the late 1970s, it remains relevant as attested by a recent empirical validation (Pearson & Shim, 1994). The names of Alter’s taxa are literally the generic names of decision support systems and each DSS taxon will contain a number of different DSS species. For example, in the representational models taxon a particular decision task may be supported either by a spreadsheet or a financial modeling package. DSS based on each technology may be regarded as different species and system evolution that involves transferring and changing the decision logic from one platform to another may be thought of as DSS speciation. Table 1. Alter’s Taxonomy of Decision Support System Taxa Description File Drawer Systems allow immediate access to data items Data Analysis Systems allow manipulation of data by tailored or general operators Analysis Information Systems provide access to a series of databases and small Accounting Models calculate the consequences of planned actions on using accounting definitions Representational Models estimate the consequence of actions without using or partially using accounting definitions Optimisation Models provide guidelines for action by generating an optimal solution Suggestion Models provide processing support for a suggested decision for a relatively structured task The theory of evolution also contributes to the understanding of the tempo of DSS evolution. Many DSS evolve in a continuous fashion, similar in some ways to that proposed by Charles Darwin. This cumulative effect of numerous small changes to an application captures the organic nature of many DSS development projects. It is the tempo of evolution addressed by Courbon et al. (1978) and Keen (1980). The major difference between biological and DSS evolution in this respect is the time scale of the evolution. Biological evolution takes tens of thousand, often millions, of years whereas DSS evolution can take place over thousands of minutes. While gradual speciation is central to Darwin’s theory the empirical evidence indicates other evolutionary tempos. Saltational change, or change by leaps, was first articulated by Aristotle. Under this thesis evolution can occur in a radical fashion, in some cases associated with the emergence of new functionality (for example, flight in birds) and frequently it creates a higher taxon. Simpson (1944) termed such rapid substantial change, quantum evolution. Some biologists debate the existence of quantum evolution and have offered gradualist neo-Darwinist explanations for the emergence of major functionality (Dawkins, 1996, chap. 4). On the other hand, Lewis (1962) produced strong evidence of quantum evolution in the plant genus Clarkia. In palaeontology, Gould & Eldredge (1977) proposed the theory of punctuated equilibria where, rather than evolving through gradual continuous change, species exist in a relatively stable state for considerable time periods and evolution is concentrated in brief periods of rapid change (some as small as 10,000 years). Evolution in the form of punctuated equilibria is also common in DSS. Keeping in mind the difference in time scale discussed above, a DSS may be static in terms of the logic and operation of the system for some months and then a rapid period of change occurs. It is clear that both gradualist and saltational theories are required to explain the empirical evidence in both biology and decision support systems. Some evolutionary lines are clearly subject to gradual speciation, while some are subject to rapid, even quantum changes over relatively short time periods. For decision support systems, the common property of each form of system evolution is that they represent a change in how the user perceives the nature of the decision process. CURRENT APPROACHES TO DSS EVOLUTION The notion that a decision support system evolves through an iterative process of systems design and use has been central to the theory of decision support systems since the inception of the field. Evolutionary development in decision support was first mentioned by Meador and Ness (1974) and Ness (1975) in their description of middle-out design. This was a response to the top-down versus bottom-up debate of the time that was concerned with the development of transaction processing systems. Courbon et al. (1978) provided the first general statement of DSS evolutionary development. In what they termed an “evolutive approach”, development processes are not implemented in a linear or even in a parallel fashion, but in continuous action cycles that involve significant user participation. As each evolutive cycle is completed the system gets closer to its final or stabilised state. Courbon argued that the evolutive cycles should be continuous and as rapid as possible as DSS exists in an environment of continuous change. USER middle-out design user learning personalised use SYSTEM facilitates implementation pressure for evolution ANALYST evolution of system function Figure 1. Keen’s Adaptive Design Framework Keen (1980), building on Courbon’s work, developed a model for understanding the dynamics of decision support systems evolution. His approach was termed “adaptive design”, although adaptive development is a more accurate term as the approach comprises development processes other than design. The importance of this work was to give the concept a larger audience. Keen (1980) remains the most cited and thereby the most influential description of the evolutionary approach to DSS development. Keen’s model comprises three major elements: the builder or systems analyst, the user and the system. These elements influence each other in complex ways during the development process. Three major iterative loops are identified in this model, namely the System-User, the User-Analyst and the System-Analyst loops. The model is depicted in Figure 1 with slight changes made to the terminology used by Keen. In the System-User loop, the link from the system to the user indicates that by using the system the user gains better understanding of their problems and in some way improves their decisionmaking process. The link from the user to the system indicates that the system provides support for the user’s personal needs. Ideally, the way that the system physically operates is personalised to the needs of the individual manager/user. In the User-Analyst loop, the user to systems analyst link represents the analyst learning about the user’s decision-making process and usage patterns. At the same time, the systems analyst to user link reflects the user’s discovery of the capabilities of the analyst and possibilities of the system development project. This loop assumes a much closer relationship between client and developer than in most other information systems applications. In the System-Analyst loop, the system to systems analyst link shows the pressure placed on the analyst to modify and add new functionality to the system. The systems analyst to system link, on the other hand, involves the systems analyst actually enhancing the system. The evolution of the final decision support system is a result of the cyclic operation of the various loops. Courbon (1996, p 119) describes these cycles as sequences of “action - whenever the designer implements a new version and the user works with it and ... reflection i.e. the feedback where the user and the designer think about what should be done next based on the preceding active use.” Courbon argues that these action/reflection cycles are similar to Piaget’s learning processes of accommodation and reflective abstraction (Shaffer, 1989) and concludes that DSS evolution can be best seen as a learning process. The organisational influences on DSS development can also be considered in terms of Keen’s three adaptive loops. Organisational procedures such as control, communication and reward systems may limit user discretion and behaviour. Also, users can exert pressure to change procedures which constrain organisational learning. In a similar manner, the technology available within the organisation limits the capabilities of the system. Other organisational issues include the charter and location of the development staff. Keen’s model is notable for its consideration of environmental factors on DSS evolution. Sprague and Carlson (1982) in an analysis of system adaptation and evolution identified four levels of DSS flexibility: the flexibility to solve a problem in a personal way; the flexibility to modify a specific DSS to handle different problems; the flexibility to adapt to major changes that require a new specific DSS; and the flexibility to evolve with changes in technology. They believed that these levels exist in a hierarchy with evolution at the top. They argued that “DSS must evolve or grow to reach a ‘final’ design because no one can predict or anticipate in advance what is required. The system can never be final; it must change frequently to track changes in the problem, user, and environment because these factors are inherently volatile” (Sprague and Carlson, 1982, p132). Sprague and Carlson’s ROMC design method was designed to provide this flexibility in DSS development. A number of cases have reported the use of the ROMC approach for evolutionary development (eg. Igbaria et al., 1996) . There have been numerous other contributions to the understanding of the evolution of DSS. Keen & Gambino (1983) provided an important case study of evolutionary development that was influential in the development of DSS methodologies, particularly with regard to their finding that evolution occurred at the sub-task rather than the task level. Stabell (1983) placed evolutionary development in a decision theoretic framework by suggesting that the evolution of a DSS should take place in a tension between the descriptive and prescriptive views of the target decision (Bell, Raiffa & Tversky, 1988). Stabell suggested that early in a DSS project the analyst should describe the current decision process and define an appropriate normative solution. He argued that DSS evolution should be planned along a continuum between these two definitions. Each version of the DSS should creep towards the normative solution. Young (1989) presented a three stage DSS methodology whose final stage of iterative use, refinement and assessment is an example of evolutionary development. Sage (1991, chap. 5) outlined a seven stage systems design methodology where the stages were sequenced in an iterative manner. Sage noted that information requirements determination exists in all stages of the DSS development process and that this is likely to be the driver of system evolution. Arinze (1991) developed a research model for DSS methodologies that was based on evolutionary principles. He saw DSS methodologies as a tool for reducing the “unstructuredness” of managerial decision making. Silver (1991) embraced the evolutionary philosophy in his framework for DSS research and practice. He extended evolutionary theory by considering how DSS restrict or limit decision making processes and how DSS can guide or direct a user’s approach to the operation of the system. Fitzgerald (1991) in a survey of executive information systems development practice reported that all the developers who were interviewed used an evolutionary approach. Keen’s adaptive design model has been extended to executive information systems (Suvachittanont et al., 1994). Evolutionary development has also been used in group decision support systems (Shakin, 1991). The discussion of current approaches and theories of evolutionary development illustrates the importance of the concept of evolution for DSS theory and practice. Existing work on the development of decision support systems emphasises a process whereby the final system results from an adaptive process of user/analyst learning and system change. However, in practice user/analyst learning can be seen as intermediate factor in adaptation. What is, or should, be of more interest to a systems analyst are the factors that trigger, enable or force this learning to take place. Another shortcoming of the current approach is that evolution is treated in a rather homogeneous way, stressing or assuming a gradualist approach. From the discussion of evolutionary theory it is arguable that other tempos of evolution are likely to occur. The next section addresses these two aspects of DSS theory: causes or triggers of adaptive and evolutionary processes, and the tempo or dynamics of evolution. A FRAMEWORK FOR UNDERSTANDING DSS EVOLUTION The aim of this section is to develop a framework for understanding DSS evolution that is founded on the theories and evidence discussed in the previous sections. This framework is structured around the aetiology, lineage and tempo of DSS evolution. Aetiology refers to the causes of evolution, lineage to whether evolution occurs within an application or between applications, and tempo relates to the pattern of evolution over time. The lineage that develops as a result of DSS evolution can be conceptualised at the application level; an application is a computer-based information system that supports an aspect (sometimes all) of the decision task. Evolution can be thought of as either occurring in a lineage within a type of application, or as occurring as a branching lineage between different applications. An example of within-application evolution is substantial change in the decision logic of a spreadsheet system such that the change is so great that the current system is indistinguishable from the first version. Evolution between applications is likely to have a more significant effect on systems development and will probably involve a new set of system initiation tasks, including technology selection, initial requirements analysis and application budgeting. An example of this form of evolution is a project that starts out as a data oriented system using EIS software and over time moves to one focused on complex financial modeling. Another way of viewing DSS evolution is to consider its aetiology, in particular the forces or factors that trigger the adaptive processes that lead to evolution. These factors may be exogenous or endogenous with respect to the decision maker. DSS aetiology is summarised in Table 2. Endogenous, or cognitive triggers have been emphasised in many research studies (Courbon, 1996; Keen, 1980; Keen & Gambino, 1983; Sprague & Carlson, 1982; Valusek, 1994). The most common expression of these evolutionary triggers is when managers learn more about the decision task by using the system and interacting with a systems analyst. However, other forms of cognitive trigger are possible. A manager may think of a new system requirement during a conversation with a fellow manager. The idea for a change in the logic of a system could also come from a consultant other than the DSS analyst, especially if the consultant is a domain expert rather than an information systems professional. Attendance at conferences, seminars and training courses could also provide cognitive triggers for DSS evolution. Finally, just simply thinking about the DSS and the decision task (say while driving to work) could lead to ideas that trigger evolutionary changes to the system. Table 2. The Aetiology of Decision Support Systems Evolution Cognitive Triggers Environmental Triggers System use Technology change Analyst interaction Personnel change Peer interaction Internal organisational change Consultant interaction Merger/Takeover Training courses Industry changes 'Idle' thought Coevolution Environmental triggers have not been addressed by DSS researchers to the same extent as cognitive triggers. Angehrn and Jelassi (1994) in a review of DSS research and practice saw systems evolution as necessary to cope with constant environmental change. The most obvious environmental trigger is a change in the technology available to the systems analyst. The emergence of personal computers was a major factor in the evolution of many DSS in the 1980s. Changes in data base software, particularly multi-dimensional models, and the internet have had a similar effect in the 1990s. Another environmental trigger of system evolution is a change in the user of the DSS. The new user may have a different conceptualisation of the task and different cognitive abilities, and will wish the system to reflect their personal approach. Internal organisational change may trigger evolution of the decision task level, and as a consequence, the DSS. These changes include divisional and departmental restructuring, downsizing, and outsourcing. Wholesale organisational changes such as merger and takeover can lead to a change in organisational procedures which will trigger DSS evolution. External events such as change to the industry structure and changes in government regulations may also require fundamental changes to a DSS. The final environmental trigger is coevolution, which occurs when a major change in the decision logic of one application or evolutionary line triggers change in another. Combining different lineages and aetiology yields the framework of DSS evolution presented in Figure 3. The framework identifies four major classes of DSS evolution, namely within application with a cognitive trigger, within application with an environmental trigger, between application with a cognitive trigger, and between application with an environmental trigger. As argued above, DSS evolution can follow a number of different patterns over time. DSS may evolve in a continuous fashion, be subject to punctuated equilibrium, and may exhibit quantum evolution. These tempos may be determined by aetiology and lineage. The likely allocation of these tempos to the classes of DSS evolution is shown in Figure 3. Evolution becomes more discontinuous moving from the upper left cell to the lower right cell of Figure 3. Cognitive Trigger Environmental Trigger Figure 3. Within Application Between Application Continuous Evolution Punctuated Equilibrium Continuous Evolution Punctuated Equilibrium Continuous Evolution Punctuated Equilibrium Quantum Evolution A Framework for Decision Support Systems Evolution To illustrate the descriptive validity of the framework, Figure 4 presents some examples of the different classes of DSS evolution from the research literature. It is difficult to analyse much of the reported research in evolutionary terms because many papers simply state that an evolutionary approach was used without further elaboration. Few mention the causes or triggers of system changes and even fewer mention the time frame of development. Nevertheless, one can find many illustrations of cognitive triggered within-application evolution. Igbaria et al. (1996) described the development of a DSS for transport managers in a New Zealand dairy company where the development began with a replication of current practice on the computer and further changes to the systems were mainly triggered by managerial learning. Courbon (1996) outlined a DSS that was developed to schedule nurses on various shifts over a four week period. The system did not use a normative scheduling algorithm and was modified as the nurses used the system and learnt more about the problem. Keen & Gambino (1983) described in great detail the development of a DSS to support decisions related to public school financing. This development project is a clear example of continuous evolution that is triggered by manager and analyst learning. Within Application Cognitive Trigger Environmental Trigger Figure 4. Between Application Courbon (1996) Igbaria et al. (1996) Keen & Gambino (1983) Botha et al. (1997) Hurst et al. (1983) Janson & Smith (1985) Nandhakumar (1996) Niehaus (1995) Rigby et al. (1995) Gunter & Frolick (1991) Jirachiefpattana et al. (1996) Suvachittanont et al. (1994) Examples of Decision Support Systems Evolution Examples of within-application evolution with environmental triggers are also common. Niehaus (1995) described punctuated evolution within a human resource planning application in a large naval dockyard. The triggers for the evolution of this system were improvements in technology and the major organisational change associated with restructuring and downsizing. Rigby et al. (1995) detailed a case of punctuated evolution in an oil industry application. The evolution of this DSS was enabled by the new functionality provided by advances in information technology. Nandhakumar (1996) in a detailed description of the evolution of a system to support inventory performance found that within-application evolution of the system was caused by a major change in organisational reporting procedures. The evolution of this system is best characterised as punctuated equilibrium. Examples of between-application evolution are more difficult to find in the research literature but are not necessarily less common in practice. Botha et al. (1997) presented an example of this form of evolution in a case from the South African National Defence Force. The system comprised six modeling applications and a supporting database application. Hurst et al. (1983) in their case describing a DSS to support production and finance decisions in a health and beauty products company, reported on continuous between-application evolution with a strong element of user learning triggered by system use and analyst learning. Janson and Smith (1985) reported on a case study of the development of a DSS to support marketing research and corporate planning in an insurance firm. The system began as a data analysis application and then progressed to representational models. Each aspect (application) of the system evolved through a number of versions. Gunter and Frolick (1991) described three generations of executive support at a power utility where the availability of new technology enabled new applications to be developed. Jirachiefpattana et al. (1996) in their Energy Company case found that the system underwent between-application evolution as a result of significant changes to the business of the corporation. Suvachittanont et al. (1994) in a case study of evolutionary development of an executive information system in a manufacturing company, related how between-applications evolution occurred when the development team had spare capacity. Each new application was prioritised according to the seniority of the manager requesting the new application. While these research studies have been allocated exclusively to one class of DSS evolution or another in the framework, it is likely that many projects will exhibit a number of these evolutionary classes over the various generations of the DSS. CONCLUDING COMMENTS This paper contributes to decision support systems research by further clarifying the nature of DSS evolution. The main differences between biological and DSS evolution were argued to be the time frame for evolution, the use of artificial rather than natural selection in DSS, and the subjective nature of DSS, in particular problem and generation definition. Despite these differences many of the aspects of the theory evolution are relevant to decision support systems. Using insight from a number of sciences and prior DSS research, a framework for understanding DSS evolution was developed. This framework is based on the aetiology, lineage and tempo of evolution. The aetiology of DSS evolution was discussed in terms of cognitive and environmental triggers, an important distinction that has not been common in published DSS research. The lineage of DSS evolution was considered as occurring within an application and between applications. The time pattern of evolution was viewed as a consequence of lineage and aetiology. DSS were considered capable of continuous, punctuated, or quantum evolution. The descriptive validity of the framework was demonstrated by applying it to published DSS studies. While the classes of evolution identified in the framework have obvious interdependencies, it is clear that each form of evolution places different demands on the systems analyst and in turn the user. Evolutionary development is an uncertain and often stressful process. Better theories of DSS evolution may help systems analysts predict what may happen next in the development processes and help them in deciding which techniques and tools are likely to succeed with each class of evolution. It is important to identify which class of evolution is occurring in order to effectively manage the evolutionary processes during a DSS development project. It seems that between-application evolution is more difficult to manage than within-application evolution. Further research, especially intensive case studies, is needed on these matters. The notion that DSS evolution is best viewed in terms of aetiology, lineage and tempo gives structure to this further work. In particular, further investigation of the triggers of DSS evolution is likely to be of significant practical relevance. REFERENCES Alter, S.L. (1980). Decision support systems: Current practice and continuing challenges. Addison-Wesley, Reading, Massachusetts. Angehrn, A.A., & Jelassi, T. (1994). DSS research and practice in perspective. Decision Support Systems, 12, 257-275. Arinze, O.B. (1991). A contingency model of DSS development methodology. Journal of Management Information Systems, 8 (1), 149-166. Bell, D.E., Raiffa, H. & Tversky, A. (1988). Descriptive, normative and prescriptive interactions in decision making. In D.E. Bell, H. Raiffa and A. Tversky (Eds), Decision Making: Descriptive, Normative and Prescriptive Interactions (pp 9-30). Cambridge University Press. Botha, S., Gryffenberg, I., Hofmeyer, F.R., Lausberg, J.L., Nicolay, R.P., Smit, W.J., Uys, S., van der Merwe, W.L. & Wessels, G.J. (1997). Guns or butter: Decision support for determining the size and shape of the South African National Defence Force. Interfaces, 27 (1), 7-28. Budd, A.F., & Coates, G.C. (1992). Non-progressive evolution in a clade of Cretaceous Montastraea-like corals. Paleobiology, 18, 425-446. Courbon, J-C. (1996). User-centered DSS design and implementation. In P. Humphreys, L. Bannon, A. McCosh, P. Milgliarese & J-C. Pomerol (Eds), Implementing systems for supporting management decisions: Concepts, methods and experiences (pp 108-122). Chapman and Hall, London. Courbon, J-C., Grajew, J. & Tolovi, J. (1978). Design and implementation of interactive decision support systems: An evolutive approach. Unpublished Manuscript. Institute d’Administration des Enterprises, Grenoble, France. Darwin, C. (1996). The origin of species. Oxford, UK: Oxford University Press. (Original work published 1859) Dawkins, R. (1996). Climbing mount improbable. Viking, London. Fitzgerald, G. (1991). Executive information systems and their development in the UK: A research study. International Information Systems, 1 (2), 1-35. Futuyma, D.J. (1986). Evolutionary biology. Second Edition. Sinauer, Sunderland, Massachusetts. Galliers, R.D. (1992). Choosing information systems research approaches. In R.D. Galliers (Ed.), Information systems research: Issues, methods and practical guidelines (pp. 144-162). London: Blackwell Scientific. Gould, S.J., & Eldredge, N. (1977). Punctuated equilibria: The tempo and mode of evolution reconsidered. Paleobiology, 3, 115-151. Gunter, A., & Frolick, M. (1991). The evolution of EIS at Georgia Power Company. Information Executive, 4 (4), 23-26. Hurst Jr., E.G., Ness, D.N., Gambino, T.J., & Johnson, T.H. (1983). Growing DSS: A flexible, evolutionary approach. In J.L. Bennett (Ed.), Building decision support systems (pp 111-132). Addison-Wesley, Reading, Massachusetts. Igbaria, M., Sprague Jr., R.H., Basnet, C. & L. Foulds, L. (1996). The impacts and benefits of a DSS: The case of FleetManager. Information and Management, 31, 215-225. Janson, M.A., & Smith, L.D. (1985). Prototyping for systems development: A critical appraisal. MIS Quarterly, December, 305-315. Jirachiefpattana, W., Arnott , D.R., & O’Donnell, P.A. (1996). Executive information systems development in Thailand, In P. Humphreys, L. Bannon, A. McCosh, P . Milgliarese & J-C. Pomerol (Eds), Implementing systems for supporting management decisions: Concepts, methods and experiences (pp 203-224). Chapman and Hall, London. Keen, P.G.W. (1980). Decision support systems: A research perspective. Data Base, 12 (1/2), 15-25. Keen, P.G.W., & Gambino, T.J. (1983). Building a decision support system: The mythical man-month revisited. In J.L. Bennett (Ed.), Building decision support systems (pp. 133-172). Reading, MA: Addison-Wesley. Lewis, H. (1962). Catastrophic selection as a factor in speciation. Evolution, 16, 257-271. Loewenstein, G. (1996). Out of control: Visceral influences on behaviour. Organisational Behaviour and Human Decision Processes, 65 (3), 272- 292. Meador, C.L., & Ness, D.N. (1974). Decision support systems: An application to corporate planning. Sloan Management Review, 15 (2), 51-68. Nandhakumar, J. (1996). Design for success?: Critical success factors in executive information systems development. European Journal of Information Systems, 5, 62-72. Ness, D.N. (1975). Interactive systems: Theories of design. Joint Wharton/ONR Conference on Interactive Information and DSS, The Wharton School, University of Pennsylvannia, Philadelphia. Niehaus, R.J. (1995). The evolution of strategy and structure of a human resource planning DSS application. Decision Support Systems, 14, 187-204. Pearson, J.M. & Shim, J.P. (1994). An empirical investigation into decision support systems capabilities: A proposed taxonomy. Information and Management, 27, 45-57. Price, P.W. (1996). Biological evolution. Harcourt Brace, Orlando, Florida. Rigby, B., Lasdon, L.S., & Waren, A.D. (1995). The evolution of Texaco’s blending systems: From OMEGA to StarBlend. Interfaces, 25 (5), 64-83. Sage, A.P. (1991). Decision support systems engineering. John Wiley & Sons, New York. Shaffer, D.R. (1985). Developmental psychology: Theory, research and applications. Brooks/Cole, Monterey, California. Shakun, M.F. (1991). Airline buyout: evolutionary systems design and problem restructuring in group decision and negotiation. Management Science, 37 (10), 1291-1303. Silver, M. S. (1991). Systems that support decision makers: Description and analysis. Wiley, Surrey, UK. Simpson, G.G. (1944). Tempo and mode in evolution. Columbia University Press, New York. Sprague Jr., R.H., & Carlson, E.D. (1982). Building effective decision support systems. Prentice-Hall, Englewood Cliffs, New Jersey. Stabell, C.R. (1983). A decision oriented approach to building DSS. In J.L. Bennett (Ed.), Building decision support systems (pp 221-260). Addison-Wesley, Reading, Massachusetts. Suvachittanont, W., Arnott, D.R., & O’Donnell, P.A. (1994). Adaptive design in executive information systems development: A manufacturing case study. Journal of Decision Systems, 3 (4), 277-299. Valusek, J.S. (1994). Adaptive design of DSS: A user perspective. In P. Gray (Ed.), Decision support and executive information systems (pp 78-86). Prentice Hall, Englewood Cliffs, New Jersey. Young, L.F. (1989). Decision support and idea processing systems. William Brown, Dubuque, Iowa. COPYRIGHT David R. Arnott (c) 1998. The author assigns to ACIS and educational and non-profit institutions a non-exclusive licence to use this document for personal use and in courses of instruction provided that the article is used in full and this copyright statement is reproduced. The author also grants a non-exclusive licence to ACIS to publish this document in full in the Conference Papers and Proceedings. Those documents may be published on the World Wide Web, CD-ROM, in printed form, and on mirror sites on the World Wide Web. Any other usage is prohibited without the express permission of the author.