I had the opportunity to pick up The Mythical Man Month the other day when I was visiting the local Borders and I must say that this book literally drips truth. Many of the essays resonated with me at a very deep level, oftentimes pointing out reasons for pain in software projects that I had experienced but never really thought through previously.
This is not really a book, but a collection of short essays, each readable easily in a single sitting (even with a compressed schedule like mine). Being the 20th aniversary some of the examples speak of systems and languages before my time, but the great thing about our field is that the fundamentals rarely change. I found the examples to be engaging, the concepts pointed out to be mostly accurate (even after 20 years), and can only say that this should be required reading both for software developers who have any aspirations of management or becoming and architect and particularily anyone who manages software projects.
What sort of truths are in there? Well, I don't want to spoil it for you, so pick up the book, but things like "Adding more workers to a late software project just makes it later" and that good systems design needs to come from a small group and implemented with a consistant vision are two concepts that are often forgotten or outright ignored in our field.