AQ's answer covers very well what I would consider to be the philosophy of computer science. Without an understanding of this philosophy, you will spend a LOT of time rediscovering things that are already known.
However, there are other types of books. Some of these can be more important from a strictly practical standpoint. Even if you don't read the actual book, you should have an understanding (and experience with) the material.
(Note: many of these are covered in typical computer science and electrical engineering courses, as well. You may just look up the course syllabi at your local/prestigious university.)
Fundamentals of computer science:
A data structures book (I like this one, but it's because Michael Main was an amazing teacher)
A programming languages fundamentals book that gives at least a high level understanding of compilers. I'd recommend one, but I've not yet found a great one.
An introductory algorithms book, like... Introduction to Algorithms. No O(x^n) code! And if you don't understand that, it's why you need the book
Fundamentals of electrical engineering & construction: At this point, I admit that I'm a comp sci kinda guy, so I'm shallow here, but...
A book about the fundamentals of computer hardware. Code by Charles Petzold seems to fill the gap pretty painlessly... not necessarily rigorously, but painlessly.
The electrical engineers I work with swear that Practical Electronics for Inventors is absolute gold. It does things "the right way, finally!".
Unlike traditional programmers, we often need to actually make non-product level stuff in the physical world. I always recommend Building Scientific Apparatus, and a mentor to help out through the rough spots.
ni.com is no slouch in this department... their white papers are occasionally (obviously, irritatingly) marketing tracts, but more often contain a lot of good, fundamental measurement information.
Software construction:
stackoverflow.com has pretty exhaustive lists of this type of book, but there are a couple of standouts.
Code Complete by Steve McConnell. Yes, you should read this. Once a year.
The Pragmatic Programmer by Hunt and Thomas. Not my favorite writing style, but, when you start putting it into practice, you realize how useful it really is.
Business of software: There's been a lot of software navel-gazing over the years. That said, a lot of it is just useful navel-gazing. And all of us do work in software, or want to... right?
Joel on Software. Joel can be irritatingly smug, on occasion. That said, his writings have an amazing amount of insight into how the software world should be, and sometimes isn't. And hey, he has a book list, too. That looks familiar...
Peopleware. Once a year. Twice if you're a manager. Not kidding.
I really like randsinrepose.com. I suspect it's just a personal taste, but there it is.
Probably the most important thing is... READ! Something, anything! Most programmers (and test engineers) don't. And it gives you a huge advantage when you do.
Joe Z.