Originally sought out after a recommendation from Hillel:

I’m reading Nancy Leveson’s book and am still thinking about it, but one neat thing: a lot of programmers love to explain why software engineering is different from ‘conventional engineering’. Leveson is both kinds and her reasons are completely different from the usual claims. […]

A lot of people who talk about why ‘software is different from other engineering fields’ haven’t actually done other kinds of engineering, so their claims aren’t based on personal experience. To know the differences, we need to talk to the crossovers.

The book is about a new approach to safety engineering: treat safety as an emergent property of a collection of interrelated components. Safety is a set of constraints on the design of the system, and accidents occur when those constraints aren’t properly controlled.

It’s a good model, and explained thoroughly and in detail. I only read the first half of the book (it’s quite dense); what I read I found interesting. I expect I’ll revisit the second half at some point, possibly skimming rather than reading every word.

(see all reviews)