How Complex Systems (don't) Fail
Are successes because of the design or despite the design? What about failures?
Dr. Richard Cook of the Cognitive Technologies Lab has been researching complex systems for 25+ years. He published a short pdf 15 years ago, called How Complex Systems Fail or "Short Treatise on the Nature of Failure; How Failure is Evaluated; How Failure is Attributed to Proximate Cause; and the Resulting New Understanding of Patient Safety”. He presents a series of 18 points on how complex systems are designed (how they emerge?) and where they typically fail. I like his point that there are rarely true single failures in these systems - rather it is often a collection of small failures or changes that end up creating the circumstances for the failure to happen. Pinning the cause on one event (or a person) is often the wrong answer.
Interestingly, the work I have been doing in TOC takes a different view of systems and the forces that drive them. Of course, TOC isn’t usually looking at failures per se, but rather a system that isn’t working as well as it could. But in a system with many moving parts, the changes need to be thought through in terms of potentially negative ramifications. Or “the system” will resist the change because it has been designed - explicitly or implicitly - to work under the old way of doing things. And, of course, TOC has a deep belief in inherent simplicity of systems, rather than the default assumption that everything is complex. But these are different ideas of complex.
He also presented at the O’Reilly Velocity 2012 conference on the same topic, though he introduced it as “How Complex Systems Don’t Fail” - or why they don’t fail as often as it seems like they should. This was recorded as a ~30 minute video. He covers some of the same ground that is in the paper, but he also brings out the ideas of resilience and the engineering that does into designing systems.
Previous entry: Critical Chain and Thinking Process videos