This is THE book on managing requirements in Unified Process.This is certainly not the book of choice about how to manage requirements as a business analyst, for instance. However, this text presents a very broad range of tools and practices in this field. It's even probably the most comprehensive one that I know. Of course each topic can't be handled in depth, but each one is pretty well presented and it's abious that the author master his subject very well. It would be not my book of choice on requirements management, but it's a very good introduction to each and every practice. The reading is relevant even outside the context of Unified Process.
Lefingwell describes requirements analysis and specification in great detail in this book. He tackles the process of understanding the problem, stakeholders, and coming up with a solution in a step by step guide (more or less), while addressing some of the common issues of eliciting and specifying requirements.
The author did tackle the issue of choosing the correct requirement methodology, and he did address the agile method. However, I do not believe that they were given justice. The book spends most of its chapters explaining the steps to produce an SRS document but this does not reflect most of the current industry standards, as most of them follow one of the agile methodologies and have left the traditional methods behind.
In the end, I would recommend this book to someone who is new to the software engineering field. But I don't think it holds a lot of information for a veteran.
The was one textbook I actually read completely in college. It was absolutely terrible. Its points were obvious. Its case study was imaginary so it wasn't really a case study. This is a poor model for a textbook.