Normal Accidents: Living with High-Risk Technologies

This book introduces a theory of complex systems and why they may be more accident prone than non-complex ("linear") systems. Perrow then explores how the theory plays out in multiple industries, which all vary in interesting ways: nuclear power, chemical plants/refineries, air travel, marine transport, mining/dams, and more.I work in high tech (large scale online systems, think Big Tech internet/cloud computing) and the theory described in this book is extremely prescient in anticipating the sort of technical problems we sometimes encounter today -- despite having been written 40 years ago. I wish I had read this years ago, to be honest.My one minor complaint with the book: the 1999 reprint, which includes a pretty hastily assembled and speculative Y2K treatment (written before the fact) features a photo of the Challenger accident on the cover. In fact the Challenger is not a good example of a system accident (it was just a bad component design, plus institutional failure/pressure to keep flying it). A much better example would be a nuclear meltdown (e.g. Chernobyl), which Perrow predicted multiple times; again his text being written in 1982-1983 mainly. Anyhow, the publisher can be forgiven for insufficiently thought-through cover photo selection.

Perrow provides an insightful framework for understanding the complex systems we live with every day and the ways they fail. Unfortunately, he spends his conclusion trying to make policy recommendations that aren't actually well supported by his own framework.I highly recommend nearly all of this book for anyone that will be designing, operating, or criticizing complex systems. Aside from poorly explained comments on nuclear criticality accidents and naval nuclear reactors, the earlier chapters of the book are technically quite sound, and Perrow's framework is a good starting point to think about how to make complex systems as safe and resilient as possible.

Well written and edited.Really highlights just what a flimsy house of cards we have built.Worse yet, most of the "solutions* end up making things less safe.A must read for any Engineer.

I have been mulling over this review for a while now, and am still undecided on the correct rating to award this book. On the one hand Perrow offers some genuine insight into systems safety, but frequently does not understand the technicalities of the systems (or occasionally their operators) well enough to make informed decisions and recommendations. In more egregious cases he comes to conclusions that are guaranteed to reduce safety (as when he argues that supertankers should be run by committee, and the usefulness of the Captain is no more) or are merely the cherished liberal opinions of an Ivy League sociologist (he teaches at Yale) as when he argues for unilateral nuclear disarmament, government guaranteed income plans, and heroin maintenance (distribution) plans for addicts "to reduce crime." In the case of disarmament, remember this was written during the early 1980s while the Soviet Union was still a huge threat...complete nuclear disarmament would have resulted in fewer US nuclear accidents, but would NOT have made us safer as we would have been totally vulnerable to intentional nuclear attack. He has great personal animosity toward Ronald Reagan, and makes inflammatory statements in the mining section that mining safety regulations would surely be weakened by Reagan, causing many more accidents and deaths. Later in the same section, though, he concludes that mining is inherently dangerous, and no amount of regulation can make it safe. So which is it? Any of this is, at very best, folly, but regardless of political bent (he is a self avowed "leftist liberal") has absolutely no place in a book ostensibly on safety systems. As such I think portions of this book show what is so wrong in American academia today: even genuinely excellent research can be easily spoiled when the conclusions are known before the research is started. This is one of the many reasons that physical scientists scorn the social sciences, and it doesn't have to be this way.Having said all that there IS a wealth of good information and insight in this book when Perrow sticks to systems and their interactions. The book contains the finest analysis commercially available of the Three Mile Island near-disaster, and his insight about how to improve safety in nuclear plants was timely when the book was written in 1984, though many improvements have been made since then.Speaking as a commercial airline pilot, I feel his conclusions and observations about aircraft safety were generally true at the time of printing in 1984, but now are miserably out of date. (The same is true of the Air Traffic Control section.) I believe that he generally has a good layman's grasp of aviation, so I am willing to take it as a given that he has a knowledgeable layman's comprehension of the other systems discussed. As an aside, he never gets some of the technicalities quite right. For instance, he constantly uses the term 'coupling' incorrectly in the engineering sense; this is particularly objectionable in the aviation system where it has a very specific meaning to aeronautical engineers and pilots.The section on maritime accidents and safety is superbly written. Here I am not an expert, but there seems to be a high degree of correlation with the aviation section. His section on "Non Collision Course Collisions" by itself makes this book a worthwhile read. He presents very compelling information and reasoning until the very end of the section, at which point he suggests that since ships are now so big, large ships (especially supertankers) essentially should have no Captain, but should be run by committee. This is an invalid conclusion, and he offers no evidence or substantial argument to support that idea. Clearly, it is an idea hatched in his office and not on a ship (or plane.) There always needs to be a person in a place of ultimate authority in fast moving, dynamic systems, or the potential exists to have crew members begin to work at direct odds with each other, making a marginal situation dangerous. Ironically, in the very same part of the discussion where he concludes that there should be no Captain, he has hit upon the key to the problem. He mentions that he was pleased to see that some European shippers were now training their crews together as a team, and that he expected this to lower accident rates. He is, in fact, exactly right about that. Airlines now have to train crews in Crew Resource Management (CRM) in which each member of the crew has the right and obligation to speak up if they notice anything awry in the operation of their aircraft, and the Captain makes it a priority to listen to the input of others, as everyone has a different set of concerns and knowledge. In this way, the Captain becomes much less dictatorial, and becomes more of a final decision maker after everyone has had their say. It IS critical, though, to maintain someone in command, as there is no time to assemble a staff meeting when a ship is about to run aground, or a mid-air collision is about to occur. Many other well documented studies and books have come to this conclusion, and in the airline industry since CRM was introduced the accident rate has decreased dramatically.Overall, if you have a desire to understand high risk systems, this book has a lot of good information in it; however it is woefully out of date and for that reason among others, I can only recommend it with reservations. A better and much more contemporary introductory book on the subject is 'Inviting Disaster' by James R. Chiles. Remember, this book was written over twenty years ago, and much has changed since then. There is knowledge to be gleaned here, but you have to be prepared to sort the wheat from the chaff.

The material is a bit dated, but the lessons are timeless. Perrow references an old adage that "Man's reach always exceeds his grasp". That seems to be so true when we look at today's catastrophes such as the Gulf oil spill or the financial meltdown. Perrow presents a wide variety of remarkably well researched accidents across many industries to illustrate the problems with our complex modern systems. He makes reasonable recommendations as to how to deal with these problems, and along the way, he points out the major flaws of accident investigation boards and the risk assessment profession. Read this book! Its both instructional and (morbidly) entertaining.

I have been part of the Electronics Technician FAA workforce since 1975. With that in mind I thought the Aviation and Space treatment was good in variety and detail presented.The ideas presented throughout the book of complexity and coupling make a lot of sense. The drawback to this book is that it was written in the 80's and updated in 1999. I wish I remembered more of some accidents described beyond the headlines. Some of the accidents are great examples of how analysis both in depth and quick at a glance still miss the big picture.I plan to read the Author's follow on book recently published (The Next Catastrophe...)

I'm a safety engineer, so this stuff is my bread and butter. I can tell you that this is a highly respected book in the field. A little dated now so don't take everything it says as gospel but it's essential reading to understand where modern safety systems came from.

A fine old book. Glad I could connect with a great quality used copy. It will make you think in different ways.

Great read and as valid today as it was in 1979, Printing quality in the book is poor especially for the price.

- analisi interessante -a -b -c -d -e -f -g -h -i -l m-n -o -p -q -r s-t u-v -z

かつて、組織社会学でベストセラーだった作家の作品。東電の原発事故・JR西日本福知山線脱線事故を予見したようなもの。これを読むと、首都圏での暮らし、仕事を続けることが辛くなる。さらに、身の丈にあわなかったトヨタの急成長、そして過度に近代化しようと努力した名古屋の都市機能の限界もわかる。タイ・中国に進出してしまった日本企業は後悔することだろう。オリンピック招致で浮き足立っている首都圏にある企業・学校・公共施設・交通網に従事する人は、必読。さらに、今後、地震被害が予想される中京・関西・四国沿岸部での生活、企業活動をしている人達も「誤った」備えをしないためにも、読むべきだ。