Ashgate Publishing. - Drift into failure: from hunting broken components to understanding complex systems

I want to thank Paul Cilliers and Jannie Hofmeyr at the Centre for Studies in Complexity at the University of Stellenbosch, South Africa, for our fascinating discussions on complexity, ethics and system failure. I also want to thank Eric Wahren and Darrell Horn for their studious reading of earlier drafts and their helpful comments for improvement.

"'Accidents come from relationships, not broken parts." Sidney Dekker's meticulously researched and engagingly written Drift into Failure: From Hunting Broken Parts to Understanding Complex Systems explains complex system failures and offers practical recommendations for their investigation and prevention from the combined perspectives of unruly technology, complexity theory, and post-Newtonian analysis. A valuable source book for anyone responsible for, or interested in, Organizational safety.'

Steven P. Bezman, Aviation Safety Researcher

'Dekker's book challenges the current prevalent notions about accident causation and system safety. He argues that even now, what profess to be systemic approaches to explaining accidents are still caught within a limited framework of 'cause and effect' thinking, with its origins in the work of Descartes and Newton. Instead, Dekker draws his inspiration from the science of complexity and theorises how seemingly reasonable actions at a local lavel may promulgate and proliferate in unseen (and unknowable) ways until finally some apparent system "failure " occurs. The book is liberally illustrated with detailed ease studies to articulate these ideas. As with all Dekker's books, the text walks a fine line between making a persuasive argument and provoking an argument. Love it or hate it, you can't ignore it.'

Don Harris, HFI Solutions Ltd

'Dekker's book contributes to the growing debate around the nature of retrospective investigations of safety-critical situations in complex systems. Both provocative and insightful, the author shines a powerful light on the severe limits of traditional linear approaches. His call for a diversity of voices and narratives, to deepen our understanding of accidents, will be welcomed in healthcare. Dekker's proposal that we shift from going "down and in " to "up and out" suggests a paradigm shift in accident investigation.'

Rob Robson, Healthcare System Safety and Accountability, Canada

'Professor Dekker explodes the myth that complex economic, technological and environmental failures can be investigated by approaches fossilized in linear, Newtonian-Cartesian logic. Today nearly 7 billion people unconsciously reshape themselves, their organizations, and societies through the use of rapidly-evolving, proliferating and miniaturizing technologies powered by programs that supersede the intellectual grasp of their developers. Serious proponents of the next high reliability organizations would do well to absorb Drift into Failure.'

Jerry Poje, Founding Board Member of the U.S. Chemical Safety and Hazard Investigation Board

'Today, catastrophic accidents resulting from failure of simple components confound industry. In Drift into Failure, Dekker shows how reductionist analysis breaking the system down until we find the "broken part" does not explain why accidents in complex systems occur. Dekker introduces the systems approach. Reductionism delivers an inventory of broken parts; Dekker's book offers a genuine possibility of future prevention. The systems approach may allow us to Drift into Success.'

John O'Meara, HAZOZ

When I was in graduate school for my doctorate, we always talked about the systems we studied as complex and dynamic. Aviation, nuclear power, medicine, process control these were the industries that we were interested in, and that seemed to defy simple, linear modeling industries that demand of us, researchers, safety analysts, a commitment to penetrate the elaborate, intricate and live ways in which their work ebbs and flows, in which human expertise is applied, how organizational, economic and political forces suffuse and Constrain their functioning over time.

Back then, and during most of my work in the years since, I have not encountered many models that are complex or dynamic. Instead, they are mostly simple and static. Granted, models are models for a reason: they are abstractions, simplifications, or perhaps no more than hopes, projections. Were a perfect model possible, one that completely and accurately represented the dynamics and complexity of its object, then its very specificity would defeat the purpose of modeling.

So models always make sacrifices of some kind. The question, though, is whether our models sacrifice inconsequential aspects of the worlds we wish to understand and control, or vital aspects.

During my first quarter in graduate school I took five classes, thinking this would be no problem. Well, actually, I didn't really think about it much at all. What concerned me was that I wanted as much value for money as I could get. I paid for my first quarter in grad school myself, which, for an international student, was a significant outlay (from there on I became a Graduate Research Assistant and the tuition was waived, otherwise I would not be writing this or much of anything else). For that first quarter, the earnings from a summer consulting job were burnt in one invoice. Then something interesting happened. Somehow I found out that with the four classes I had to take, I had reached a kind of maximum level beyond which apparently not even Ohio State could morally muster to extort more money from its international students. I could, in other words, throw in a class for the fun of it.