The Mythical Man-Month Anniversary Edition, Frederick P Brooks Jr, Addison-Wesley 1995, ISBN 0-201-83595-9.
This is a classic book about the pitfalls of large software programming projects ("programming-in-the-large"). Brooks wrote the first edition in 1975, after his time as project manager for IBM's OS/360 operating system. OS/360 was, for its time (1960s), a massive effort, with planned features way beyond anything available in most other operating systems of the time. Naturally it fell behind schedule and went over budget. In an effort to remedy the problems, more programmers were brought in to help, but for some reason they only made things worse. Hence Brooks' Law (page 25):
The original book also looked at other large software projects, and studies thereof, to try to derive some lessons as to how to manage these.
Later, in 1986, Brooks published an essay, No Silver Bullet, looking back at what had been learned about programming-in-the-large since the book first came out. He observed that there had been an order-of-magnitude improvement in programmer productivity over the prior ten years. This was mostly due to the shift from dependence on AssemblyLanguage towards heavier use of higher-level ProgrammingLanguages, as well as improvement in editing/debugging tools and increasing prevalence of interactive computer systems.
In the essay, Brooks also looked at the prospects for another order-of-magnitude improvement for programmer productivity in the subsequent decade. He considered a number of different technologies that various people were promoting at the time, including ObjectOrientation. His conclusion was that no single one of these technologies would offer up such an improvement in productivity. He was villified as a pessimist at the time by the proponents of those technologies, for coming to this conclusion.
Then, in 1995, this Anniversary Edition of the book was released. The original text of the 1975 book was included unchanged. To this was appended the 1986 essay, followed by another chapter called "No Silver Bullet" Refired, plus a careful listing of all the propositions made in the chapters of the original book, and notes on whether they still held true or not.
One important conclusion reached in the 1995 addenda was that there had been no order-of-magnitude improvement in programmer productivity over the prior decade. Whereas in the 1986 essay, Brooks had left open the possibility that more than one of the technologies mentioned, working in concert, might have been able to effect such an improvement, as it turned out none of them were able to achieve any such thing. Far from being a pessimist in his original conclusion, he had in fact been an optimist.