Systems Engineering Philosophy: No Easy Answers?

Transcription

1 Systems Engineering Philosophy: No Easy Answers? Indrajeet Dixit University of Southern California Los Angeles, CA Abstract Jo Ann Lane University of Southern California Los Angeles, CA Many continue to ask if Systems Engineering is really a discipline and does it have an underlying philosophy. This paper is an attempt to fill the gap in the literature on the philosophical drivers of systems engineering. In doing so, we discuss the philosophy of science, philosophy of technology, and their relationship to systems engineering. We also explore some topical issues relevant to this discussion such as, what is systems engineering, and how can it be better bounded. Our analysis reveals while systems engineering is important to the successful development and evolution of systems, it has some foundational issues resulting in an over-focus on current systems engineering practices. To better advance Systems Engineering, it is important that our research efforts expand to include the overall philosophies that support systems engineering and that we more broadly research systems engineering to include multidisciplinary principles and properties as well as innovation. Systems engineering is a new piece in the academic architecture. As compared to the older engineering specialties with their reliance on the natural sciences, systems engineering has roots in a variety of sciences, technologies and to an extent in management practices. As shown in Figure 1, it is through a variety of sciences that principles and physical properties are discovered and matured. Through the maturation process of principles and properties, technology concepts often emerge and evolve through innovation. As technologies are used, integrated into, or migrated into systems, standard practices and methodologies evolve to capture successful approaches. And finally, as the practices and methodologies mature, tools can be developed to automate tedious, complex, or error Introduction

2 prone activities. For example, through the science of physics we learn principles about gravity and through the sciences of biology and physiology we learn the principles that explain why birds can fly and man can t. Then, through the science of aerodynamics, we learn that there are ways we can make things more air-worthy. Innovation integrates several (possibly disparate) technologies, and one can develop the concept of an airplane such as the first one successfully flown by the Wright brothers. As we learned more and more about flying and airplanes, standard practices and methodologies were developed along with tools to support the development, test, and manufacture of new aircraft. Systems Engineering, in our opinion is a process enabling concept exploration as well as the development, manufacturing and testing of complex engineered systems. A recent paper (Brown, 2009) observed systems engineering researchers placed the proverbial cart (methodology) before the horse (philosophy). It was suggested that an increase in academic systems engineering research (SER) has raised concerns about rigor and validity and may have led to a greater emphasis on methodology. In addition, SER has extended from the physical and informational domains to include the social domain. Such an expansion has made philosophy paramount, yet perhaps, researchers have not expended sufficient resources identifying philosophical traditions relevant to systems engineering research. We concur with Brown, insofar as systems engineering as it is known in its modern form, has a strong social dimension. To quote Sage (pp. 164), systems engineering is a management technology which involves the interactions of science, the organization, and the environment (Sage, 2000). The inclusion of the words, management, organization and environment, lends credence to the possibility of using social science tools in systems engineering research. Subsequently, internalizing and familiarizing with the philosophy of science is essential for conducting research in the social sciences, and hence in systems engineering. Moreover, with complex organizational and institutional issues and emergent behaviours related to systems of systems 1 lurking around the corner, systems engineering scholars can ignore philosophy of science at their own peril and certainly at the cost of rigor and validity as complexity grows. This paper takes a further look at systems engineering and attempts to answer the following questions: 1. What are the philosophical roots (heritage) of systems engineering? 2. Can systems engineering be considered as a discipline or is it just a process? 3. If it is a discipline, what is the scope of that discipline? 4. What are the impacts of this with respect to future systems engineering research? Systems Engineering as a Discipline? In contrast to Brown, in our opinion, the boundary of a systems engineering problem remains ill-defined. Currently, it is difficult, if not impossible to describe, what is and what is not a systems engineering problem. While adding a social dimension does not compound the problem, it does not necessarily help bound a problem within the confines of systems engineering 2. The 1 This also raises the issue of systems of systems engineering being a distinct discipline from systems engineering. However, that is a story for another day. 2 Irrespective of the scope of a systems engineering problem, in our analysis, the social dimension remains an important lens to view any systems development activity.

3 body of literature on this particular issue while not being extensive is certainly enlightening. For example, Johnson remains unconvinced about the academic potential of systems engineering, while acknowledging its pragmatic utility (Johnson, 1997), whereas Emes et al. are unconvinced about distinguishing systems engineering from competing systems disciplines (Emes, Smith, & Cowper, 2005) and lastly, Dixit and Valerdi (2007), claim systems engineering is unable to identify problems unique to its jurisdiction and this has serious implications for the professionalization effort (Dixit & Valerdi, 2007). While these existential questions are of enormous importance, they are beyond the scope of this analysis. Our approach is conservative and pragmatic as we recognize the development of complex systems (particularly ones strongly influenced by, and hypothetically capable of influencing, institutional, regulatory and organizational environments) remains a matter of serious inquiry, requiring multiple disciplines for developing theories and testable hypotheses and consequently (if possible) repeatable solutions. Systems Engineering 3 is at the intersection of engineering and social science 4. Approaching the issue of philosophy from the perspective of philosophy of science, is necessary but not sufficient. In our opinion, there is an extant philosophical tradition, the philosophy of technology, which researchers must draw upon, in addition to the philosophy of science. The Philosophy of Everything? From a systems engineering perspective, there are two central issues: firstly what is the distinction between the two philosophies (of science and technology), and secondly, what is their relevance (if any) to systems engineering research. The definition of the key terms involved here are as follows (Merriam-Webster, 1998). Philosophy: the most basic beliefs, concepts, and attitudes of an individual or group Science: knowledge or a system of knowledge covering general truths or the operation of general laws especially as obtained and tested through scientific method Technology: the practical application of knowledge especially in a particular area Methodology: a body of methods, rules, and postulates employed by a discipline : a particular procedure or set of procedures 3 It is difficult to avoid the debate over systems engineering and engineering systems. Our position remains firm. Systems engineering is a process, and engineering systems is a field of inquiring, of which systems engineering is but one branch. 4 Management is excluded, because, modern management theory draws upon three distinct social science disciplines for the three areas of inquiry; economics for strategic management, sociology for organization theory and psychology for organizational behavior while studying phenomena at different levels of analyses; verticals (strategic management), firms (organization theory) and groups, teams or individuals (organizational behavior). These disciplines do not exclusively inform one area of inquiry, e.g. perspectives on organization theory are economic and sociological. The behavior of individuals within firms is informed by economics and psychology and strategic management is influenced by psychological and economic theories. For a discussion on the existential crisis in management, the following is recommended: Khurana, R From Higher Aims to Hired Hands: The Social Transformation of American Business Schools and the Unfulfilled Promise of Management as a Profession.. Princeton, NJ: Princeton University Press.

4 Philosophy of science is the study of the underlying assumptions, methods and practices of scientists. A typical starting point for discussions on philosophy of science is logical positivism. With logical positivism, there was no limit to the number of verifications required for a theory to be accepted as a universal truth. This was a significant impediment and Carnap modified it by suggesting any theory validated with empirical proof was tentatively accepted until the emergence of evidence to the contrary in which case it would be rejected. This form was called logical empiricism, and while problematic in its own right, has sustained the test of time (problem of induction, and of a priori knowledge). Karl Popper found a way around by introducing falsificationism. According to Popper, only when an observed phenomenon was contradicting established theory, would the need for additional experimentation arise. If in these succeeding experiments, the theory was falsified, then it was roundly rejected. Moreover even a single instance of falsification disproved a theory. The problems with falsificationism as a paradigm were similar to logical empiricism measurement errors and observational bias. Moreover, as Kuhn noted, it was impossible to refute an established framework on the basis of a single observation. Also, the scientific community had advanced in spite of contrary empirical evidence (e.g. oxidation, Copernican astronomy etc). In the classic work The Structure of Scientific Revolutions, Kuhn introduced the concept of scientific revolutions. According to this theory, science advances through paradigmatic shifts or revolutions, wherein on occasion, the objective of science may not be to seek the truth but instead, advancement for its own sake. Kuhn suggests that due to paradigmatic differences, rational argument is not sufficient to convert a scientist from one paradigm to another as it involves too great a conceptual leap. In other words, perhaps scientists remain in the silos of their paradigms. The criticism to Kuhn comes in two forms firstly Kuhn argues that the question of theory selection is a matter of faith. Critics have found contrary evidence and suggest the irrationality of such behavior. Secondly, for a number of social sciences (and even in natural science) the existence of competing paradigms has been found. In a more recent work, Paul Feyeraband introduced the model of epistemological anarchy. The underlying logic suggests it is necessary to violate established norms to achieve scientific progress. Conformity to rules inhibit creativity and Feyeraband argued that there are no universal rules for science, some rules maybe incommensurate with other disciplines, but they do not become unscientific. Lastly Cognitive Sociology of Science suggests science is a social construct and every scientific breakthrough involves the social world. Science is also a social enterprise and theories are appraised not just in terms of rational but also sociological criteria (Anderson, 1983; Taleb, 2007). This is by no means an exhaustive account of philosophy of science; it is a brief presentation of the key ideas and the contrarian perspectives. Philosophy of technology is a distinct field of study, and while being different, it certainly has significant commonalities. We observed that there are contrasting perspectives about the significance and the extent of study of the philosophy of technology or engineering. For example, Meijers et al. state (pp. 1): The fact that philosophers of science have not regarded technology or engineering as a subject worthy of serious study clearly emerges from well-known introductions, companions and anthologies. [Curd and Cover, 1998] and [Curd and Psillos, 2008], for example, do not have a single index entry for artifact, design, engineering or technology in 2000 pages of philosophy of science. There are some exception though, such as [Newton and Smith, 2000] which contain a small section on the philosophy of technology. (Meijers, Gabbay, Thagard, & Woods, 2009)

5 Others (Ihde, 2004; Lipton, 2010) observe that the philosophy of technology is as old as the philosophy of science, or just a little younger. However, programs in philosophy of technology are not as common as philosophy of science, and usually (if not always) the former are circumscribed by the larger number of science programs. As pointed out by Mitcham (Mitcham, 1990), there are two strands in philosophy of technology: scholars studying the internal structure of technology (pp. ix) and secondly, the more humanities oriented scholars who study the relations of technology with society. Scholars such as Brian Latour and others, focused on the latter form that is commonly known as Science and Technology Studies or STS programs. Mitcham states that the former is the engineering philosophy of technology and does not make a direct distinction between philosophy of engineering and philosophy of technology (Mitcham, 1990). The history of the philosophy of technology is rich, but as Ihde points out, the fundamental problem with it is that it still lacks a core set of problems and theories that can be used to determine its problem space (a problem not very different from the ones that systems engineering faces). Using the works of Lipton (Lipton, 2010) and Ihde (Ihde, 2004) we discuss the differences between science and technology. Bunge, in (Ihde, 2004), suggested science was ethically neutral, but technology was not, and it grappled between good and evil. Lipton did not discuss the ethical dimension but focused on differences in the drivers, types of knowledge, and outputs of scientific and engineering activity. He suggested that scientists were the drivers of science. They picked problems within their paradigm and jurisdiction and attempted to solve them. For engineers, problems were given to them either by sponsors or clients who expected them to use their knowledge, skills, and judgment for creating solutions. Scientific knowledge, according to Lipton was knowing that, with roots in propositional logic, whereas engineering knowledge was knowing how, which he considers as an ability or skill. The output of scientific activity is theory generation, whereas the ideal output of engineering activity is an artifact or a product. To understand the relevance of each of these philosophies, it is imperative to explore systems engineering. An upcoming effort (Dixit & Valerdi, 2011) analysed five different definitions of systems engineering, one by a professional society (INCOSE), two from texts (Blanchard & Fabyrcky and Sage) and two by individuals known for their contributions to systems engineering (George Friedman and Simon Ramo). Each definition provides a distinct perspective on systems engineering. They were different in terms of approach, scale, scope and end product. Hence, for the purpose of this discussion, we analyze systems using three lenses: people, processes and products. Given these parameters, we can now discuss the scope and relevance of each philosophy to systems engineering. Systems engineering research is at the intersection of the three perspectives people, processes and products. Within each perspective, there are multiple level of analyses, and a particular research problem may involve one element from one (or more) of the perspectives. The disciplines informing these perspectives encompass sciences, engineering, and social sciences. In Figure 2, we divide the three-dimensional space into three two-dimensional areas, with a particular philosophy having a more significant impact on any particular two dimensional area. For example, philosophy of technology, particularly the engineering type, is more influential in understanding the area between people and products. The area between people and processes is commonly studied by sociologists and social psychologists, hence we predict a stronger influence of philosophy of science. The last, the area between processes and products we believe is strongly influenced by the humanities philosophy of technology.

6 Figure 2. Impact of Philosophy In summary, conducting research in systems engineering is a complex task, and for epistemological and axiological reasons, it is imperative scholars focus on both the philosophy of science and the philosophy of technology as guiding lights. Current Focus Many of the systems engineering research activities today focus on the part of the discipline that is better understood and based upon past successes in systems engineering. There is a considerable amount of on-going research in the areas of engineering processes, best practices, and standards; life cycle models; agile processes; lean processes; six-sigma projects to refine processes; engineering approaches such as model-based engineering, platform-based engineering, and system of systems engineering; governance and management strategies; and security engineering. To illustrate this, we look at two example research areas: systems engineering process guidance and systems engineering cost modelling. These areas tend to focus on what to do, how to assess how well it is done, and how to estimate how much effort is required to adequate perform systems engineering. Systems Engineering Process Guidance: The INCOSE Systems Engineering Handbook is the result of a significant effort undertaken by INCOSE to capture and disseminate systems engineering best practices based on practices employed in the past on successful engineering efforts. However, it primarily focuses on the systems engineering process and does not address the question of innovation (in the 2007 edition, the word innovation appears five times in the entire document, and on no occasion does it explain how following the process may lead to innovative solutions; in fact in one case there is an explanation for over-reliance of process may lead to lack of innovation). Moreover, it does not aid in the development of new technologies or the identification or refinement of general principles or properties applicable to the development of complex engineered systems. The Software Engineering Institute s Capability Maturity Model Integrated (CMMI) has similar problems. It is not a tool for innovation, but instead is a process definition and assessment tool based upon current best practices. When closely adhered to, it focuses the engineer on doing an adequate job of engineering a system that meets requirements and has acceptable quality

7 and attempts to constrain efforts to target costs and schedules. This approach tends to inhibit or proscribe innovative talent and focuses the engineer on good enough or sufficiency. Systems Engineering Cost Modeling: Taking another leaf from the software engineering world, in the first half of the previous decade, a concerted effort was put into developing a systems engineering cost model. This model was called COSYSMO (Valerdi, 2005) and it predicts the number of person-months required for completing systems engineering activities as defined by a standard process model. However, COSYSMO s predictions are based on the availability of certain data (e.g. size drivers such as number of algorithms and number of systems requirements). Since COSYSMO starts with number of nominal (sea-level) systems engineering requirements as its size driver, it cannot be inclusive of the top level conceptual development, new technology development/maturation, and innovation aspects of systems engineering (since that work must have already been completed to arrive at the sea-level requirements). It is at this point, the requirements are turned over to the engineering team for further analysis, refinement, and detailed concept development. While significant engineering activity needs to be performed for a system to reach the manufacturing stage, the act of innovation has occurred long before. Inasmuch as the community benefits from the ability to predict the person-months of systems engineering activity required, there is no measure of how much innovation costs, or how long it took, or why the act of innovation occurred as it did. What s Missing: Amidst all these exciting developments, something fundamental is lacking. Reviewing its history, systems engineering was not intended to be just a bureaucratic schema, but also an engine of innovation. In his work, Rechtin calls the period of innovation (or conceptual development), systems architecting, a higher-order process of synthesis (over analysis) resulting in the creation of unprecedented systems. Lately this distinction has blurred, and architecting is wrongly equated with architecture development (the latter is the product of the former, but they are not the same). Rechtin famously theorized these acts of innovation were the product of the use of heuristics developed by an architect over years of experience (Rechtin, 1991). Hence, gaining a deeper understanding of the architecting (or innovation) process is essential in training future systems architects (who typically rise from the ranks of the larger engineering workforce). Some contemporary research such as Lane et al. delves into understanding why certain firms are consistently good innovators. The explanation lies partly in sound process techniques (shorter cycle times, excellent concurrent engineering), teamwork (small agile teams) and a culture focused on constant improvement and business interests, but also balances these activities with time to pursue and mature innovative approaches and technologies (Lane, Boehm, Bolas, Madni, & Turner, 2010). Assuming we circumscribe systems architecting within systems engineering, it is clear that it remains under-researched. Improving methods, processes and tools is necessary but not sufficient and is similar to placing the cart, before the proverbial horse (in this regard, the horse is the driver of innovation (or architecting), and the cart, i.e. methods, processes and tools or engineering ) are the enablers for this process. However, research is limited with respect to innovation, new technology development, or the development of new scientific principles/properties in the systems engineering literature this is often relegated to corporate organizations with significant financial resources trying to develop the next major gee-whiz product that will get consumers running to the store to be one of the first to own the new product. This is seen in industries such as computing (Apple, Microsoft, Google) and cellular technologies (Qualcomm and many of the computing companies mentioned before). And companies prize these research results as secret intellectual property that must not be shared.

8 Implications For any significant progress in systems engineering as a discipline, it is necessary to understand the work of systems engineers and systems architects. Philosophy of science and philosophy of technology are frameworks providing guidance and motivation to researchers interested in pursuing this path. As mentioned earlier, the disciplines capable of providing a theoretical premise for such research, are too numerous to be named. The practice, methodology, and tools of systems engineering provide a process perspective, but not a disciplinary perspective on systems engineering as a whole. Like any process, it needs to be measured, refined and managed in order for it to be improved. However, an improved process does not imply disciplinary progress and does not lead to revolutionary approaches. It means exactly what it is: an improved process. This analysis is illustrated in Figure 3. The larger concern here remains one of training and certification efforts conducted by INCOSE. Does a certified systems engineering professional qualify as an innovator or as someone who is aware of the process and has diverse experience in the different stages of developing systems? Some recent research by Davidz and Nightingale (Davidz & Nightingale, 2008), observes the key factors for enabling systems thinking in engineers. However, the question of innovation remains to be answered. Conclusions In conclusion, systems engineering is an important multi-disciplinary area of inquiry and most recent research efforts have worked to standardize and improve current systems

9 engineering processes so that we can repeatedly develop solid systems within relatively predictable cost and schedule parameters (whether the predictable cost and schedule parameters are acknowledged or acceptable is a whole other topic of inquiry). However it is imperative to focus on Systems Engineering broadly and not forget about the harder but more impactful problems related to systems engineering principles, properties, and innovation. Currently, very little research has gone into understanding the drivers of innovation in firms practising systems engineering, and we hope that changes soon. References Anderson, P. F Marketing, Scietific Progress and Scientific Method. Journal of Marketing, 47(Fall 1983): Brown, S Naivety in Systems Engineering Research: are we putting the methodological cart before the philosophical horse?, Seventh Annual Conference on Systems Engineering Research. Loughborough, UK. Davidz, H. L. & Nightingale, D. J Enabling systems thinking to accelerate the development of senior systems engineers. Systems Engineering, 11(1): Dixit, I. & Valerdi, R Challenges in the Development of Systems Engineering as a Profession?, 17th Annual INCOSE Symposium. San Diego, CA: INCOSE. Dixit, I. & Valerdi, R Some Observations on Systems Engineering as a Profession. Emes, M., Smith, A., & Cowper, D Confronting an identity crisis - How to brand systems engineering? Systems Engineering, 8(2): Ihde, D Has the Philosophy of Technology Arrived? A State of theart Review. Philosophy of Science, 71(1): Johnson, S. B Three Approaches to Big Technology: Operations Research, Systems Engineering, and Project Management. Technology and Culture, 38(4): Lane, J. A., Boehm, B., Bolas, M., Madni, A. M., & Turner, R Critical Success Factors for Rapid, Innovative Solutions, 4th International Conference on Software Processes. Paderborn, Germany. Lipton, P Engineering and Truth, Philosophy of Engineering, Vol. 1. London: The Royal Academy of Engineering. Meijers, A., Gabbay, D., Thagard, P., & Woods, J. (Eds.) Philosophy of Technology and the Engineering Sciences (1 ed.): North Holland. Merriam-Webster Merriam-Webster's Collegiate Dictionary, 10th ed.: 1559: Merriam- Webster. Mitcham, C Thinking through Technology: The Path between Engineering and Philosophy. Chicago: University Of Chicago Press. Rechtin, E Systems Architecting: Creating & Building Complex Systems. Upper Saddle River: Prentice Hall. Sage, A. P Systems engineering education. Systems, Man, and Cybernetics, Part C: Applications and Reviews, IEEE Transactions on, 30(2): Taleb, N. N The Black Swan. New York: Random House. Valerdi, R The constructive systems engineering cost model (COSYSMO). Unpublished Ph.D., University of Southern California, United States -- California.

10 Biographies Indrajeet Dixit is a Ph.D. candidate at University of Southern California. He received an M.S. in Systems Engineering from George Mason University and B.E. in Electronics from University of Mumbai (Bombay), India. His research interests span systems thinking, organizational psychology and product development processes. Dr. Jo Ann Lane is a research assistant professor at the University of Southern (USC) California Center for Systems and Software Engineering, conducting research in the areas of SoSE, systems engineering, and innovation. She was a co-author of the 2008 Department of Defense Systems Engineering Guide for Systems of Systems. Prior to her academic career, she was a key technical member at SAIC for over 20 years, responsible for the development and integration of software-intensive systems and systems of systems. She received her PhD in systems engineering from the USC and her Master s in computer science from San Diego State University.