Eli Bendersky's website "Peopleware: Productive projects and teams" seems to be the cult book on software management. So, after reading and hearing many favorable reviews I've decided to read it.
The book talks about a very familiar subject, but from an unfamiliar angle. The issue of programmer's work quality - what affects it and how to improve it, is a well-chewed topic in software engineering books. Peopleware takes a differenct side, and looks at this topic from the perspective of programmer managers and not programmers themselves.
The authors have many useful insights on the subjects of office environment (probably the most thorough treatment of this subject I've ever read), "jelling teams together", picking the right people and making work fun. I found that the 8th chapter - "You never get anything done around here between 9 and 5" has especially struck a cord with me - it was a sheer pleasure to read and to agree with.
On the bad side, I think the book re-chews certain subjects too much. Maybe an experienced manager would understand the fine distinctions, but not a mere code-monkey like me. I've read the 2nd edition, which has extra 8 chapters in the end, and found them almost completely repeating the other contents of the book. I guess the desire to earn extra money on a 2nd edition are greater than the desire to keep the book more complete and non-repeating. Also, one could not help to wonder - if these guys (Lister and DeMarco) are so smart, how come they're not big-shot managers in big firms ? They work as consultants for many years, but what is exactly the role of a consultant in this business ? Just go around firms and tell managers how they better manage ? As they say, those who can work - work. Those who can't - teach. (and those who can't teach - teach sports, but that's another issue).
All in all, I enjoyed most parts of the book. I don't think it's necessary for non-managers to read it, while managers could benefit from it. Sadly, although many things said in the book are true, they're not always easy to implement, so people just complain about these things but they eventually stay the same. I'd really like to work under a manager who not only read "Peopleware", and not only agreed with it, but also implemented at least some of the advices stated there in his team.