Of course, any major software development will have compilers, parser generators, assemblers, linkers, macro processors, text formatters, test managers, source control, custom tools that themselves need to be built, and a gazillion other bits and pieces that manage source code and its conversion to finished product. Long before you've got more than a few hundred files in a few dozen different formats and a directory hierarchy N deep, mere mortals have become unable to remember all the minutiae needed to guide every input through many steps and into the many outputs it affects. On big, active projects, even shell scripts run out of steam fast.

That's what make is for. Using a combination of pattern-matching techniques and explicit specification, it encapsulates all that knowledge of how the system is built from its pieces. Yes, it's a complex program, but it solves complex problems. If you have just a half-dozen C files and headers, this might seem like a bit much. You're right, it would just add complexity to compiling something so simple. If you've got the Unix source tree though, it's nice to know that anyone can build the whole system, with all the frills, by typing four letters: make.

You can do a lot with just a little bit of make, especially if you're porting make scripts from systems with simpler semantics (i.e. anything that's not GNU make). If you have a big system, though, you have a big problem, and you'll be happy to know about the full power that this remarkable tool puts in your hands. And, if you're thrown into a big, complicated, and existing sytem, you'll cling to the book like a life raft, at least until you learn to paddle around on your own.

Recommending this book is silly, like recommending oxygen to people who breathe. Either you genuinely need this book or you don't. If you do, then nothing else can take its place.