This is the more rather more alarmist-titled sequel to 'Why Buildings Stand Up'! Originally I was interested in this book as part of my 'get books on disasters to understand decent planning and process' thing, but in the end the failure modes for IT and buildings seem quite distinct. In buildings, the most important lesson does seem to be 'design with redundancy', which in some cases in IT is useful, but in other cases just makes things even more complex and error-prone.
So, stepping back from my personal IT angle, is this any good? Yes, it's an interesting insight into structural engineering, and really highlights how effectively ad hoc it all was before computers could be used for simulation (and presumably add their own dangers). The writing style does in places go off into flowery rubbish (generally for each chapter's conclusions), but it's mostly pretty solid. This 'updated and expanded' edition does have 9/11 shoehorned in ungracefully, but overall it makes a very pleasant and mostly orthogonal companion to '101 Things I Learned in Architecture School'.