This book is popular among the software developer crowds, analogous to a UX version of Lean Startup.
The premise of this book is that many products are designed poorly for the end-user.
This bit, 'for the end-user', is important. A poorly designed product could also be:
And so, a bad quality which should also be fairly obvious (the product sucks to use) can be masked, at least for people who don't have to use it sometimes.
Enter Don Norman. His antidote to this looks-nice-but-actually-sucks-to-use problem is 'human-centered design', which more or less amounts to 'do not make your product suck to use'.
Design of Everyday Things is not well written and is unenjoyable to read.
The poor writing is not counterweighed with intellectual heft. DoET falls into the category of a Blinkist book, that is, a book with 15 minutes of ideas spread across 200 pages of bad writing. If you buy it, buy it in an airport.
Another bother is that Norman invents his own vocabulary:
I call the signalling component of affordances 'signifiers' [page 13]
I call the combined information available to us the 'system image' [page 31]
Norman offers some rules of thumb for designers:
These rules seem pretty well understood, even by non-UX people. The book's ideas have become gospel. Should you even bother reading the book?
I'd argue 'yes', if you're in the process of designing a user-facing product. Given how often we make UX mistakes, 200 pages of the same idea could be useful repetition. You shouldn't like it, but your users might thank you.