I first read this book under a different name - Systemantics - as a photocopier-scanned PDF found online who-knows-where over ten years ago.
It made a significant impact on me, a budding software developer being entrusted to build decently sized software systems with little oversight. It instilled in me a very healthy skepticism of my own (and anyone else's) ability to design the perfect system, the perfect software architecture, even the ability to diagnose a system's flaws. It trained me to distrust anything I built until proven otherwise, to assume some amount of failure in the operation of everything built around me.
And the book - the reason I love it so much - is that it trains you to accept and cherish these aspects of the systems you build and to which you trust your life.
This second reading I wanted to do in physical form, so I bought the book from the author's widow. Recommended - she's lovely and sent me personal emails when fulfilling the order.
Because I had somet time on my hands and because I find the tone of writing entertaining, I finished this book in only six days this time around. And the structure is such that I could have highlighted something on most pages, so I won't copy excerpts here.
I kind of think that if you buy this book, it should come with a copy of Thinking in Systems - and vice versa. One will incinerate your confidence in system design and the other will rekindle your hope.
ANY LARGE SYSTEM IS GOING TO BE OPERATING MOST OF THE TIME IN FAILURE MODE.