This is a breezy book poking fun at the foibles of Unix. As a sarcastic screed, it is not at all balanced or fair or reasonable, or even necessarily historically accurate. But it is valuable.
(...)It is valuable because in many ways it is a catalog of design errors that you can make when putting together a system -- any system. Designers of new systems should be able to learn from it.
It is valuable because it shows you how over time design decisions and compromises that seemed reasonable can come to seem ridiculous.
It is valuable because it really does show you that "Worse is Better". That is, Unix really did survive, and all the 'better' systems like Multics and Tenex failed (and of course they weren't necessarily better across the board). There is a lesson here for engineers who don't understand that making the 'best' product by some narrow technical definition does NOT guarantee market success.
It is valuable because it documents some of the *alternatives* to doing things the Unix way. Not enough to substitute for studying Multics and whatever, but valuable nonetheless.
It is valuable because many of the analyses of Unix apply to other systems, certainly including MS-DOS and Windows. Yes, Windows does some things better, and some things worse. But you're smart; you can figure out how to transpose the analysis.
Finally, it is valuable because it punctures the pretensions of those who hold up Unix (and Linux) as images of perfection.
Worth reading.