I have just read "Coders at work" by Peter Seibel.
This book contained a collection of interviews with people who have developed important pieces of software such as javascript. The interviewer asks questions about the process of writing code, what books they read, how their career developed. It was interesting to see how most of them didn't use UML or any other diagram formulation, other than some pictures on a white board. Some of them had to also get other people to help them with the version control systems.
I thought this book was fantastic! However, I am not sure I would ever recommend it to students. It would distract them to see the pragmatic techniques used by these coders. Although many of the interviewers used many techniques from computer science degrees.
It would be nice to have a similar book from theoretical particle physicists, that explained what people actually did. However, given the big egos of many theorists and their hopes for Nobel prizes, many of the interviews would end up as fantasy. Personally I wouldn't want some one like Hawking to contribute.