Combinatorial algorithms Book

My opinion on this book is really divided : on the one hand I enjoy some chapters, on the other hand I hardly managed to restrain myself from flipping through other chapters. Basically, this book designs and builds a basic computer by introducing in each chapter a concept or a technology used inside computers. It was written from 1987 to 1999, consequently one shouldn't expect any description of newest technologies.

It starts really slowly with the first chapters, but then things get more and more complicated. One of the things that bother me with this book is the difference in complexity between chapters. Some chapters can be easily understood by a junior school or high school student while some of the latest chapters remind me bad memories of electronic circuits from my engineering school years. For example, a whole chapter is dedicated to explain how to communicate with your neighbour using a flashlight, an other chapter tackles the same issue with light bulbs and electrical wires, whereas all the gates or all the flip-flops are dealt with in a single chapter. I admit I have never been either fond of or good at electrokinetics, but I confess I didn't try to understand how all the electronic circuits of these later chapters work. I guess these chapters mostly interest hard code computer enthusiasts, but don't they already know these stuffs ?

Besides, few chapters are a little boring : a whole chapter to describe every op-code of Intel 8080, come on ! Does the decimal system really deserve a whole chapter ? In my opinion, decimal and alternative number systems should have been presented in a single chapter instead of two.
Moreover, the huge difference in complexity leads to some contradiction. The binary number system is so well described that a high school student can easily understand it, binary addition and subtraction are very detailed, but multiplication is done with a simple inefficient loop ! In my opinion, it would have been opportune to present at least a more efficient version based on the binary representation of the multiplicand as well as introduce exponentiation by squaring (a.k.a. square-and-multiply or binary exponentiation).
Additionally, I think that Charles Petzold tries to explain in too many details how each part works so that readers with less technical knowledge can understand, but in the end I guess these readers get lost or confused by so many details anyway, whereas a few technical references are missing. For instance, both Von Neumann and Harvard architectures are described but I don't recall them being mentioned.

Nevertheless, I really liked when the author gives historical anecdotes or references. The chapters I enjoyed the most are the ones where Charles Petzold gives readers some background history to introduce a concept or technology (for instance, Morse and Braille's codes, Bell's telegraph, the invention of telegraph relays, the evolution of transistors, chips or programming languages).

Eventually, I find it a bit contradictory for this book that most of the interesting chapters are the less technical ones indeed. Moreover, due to the important difference of knowledge required to understand chapters, I don't think someone may understand or find interesting every chapter.