A painful book on Quantum Computing, but probably a good read for one not familiar with the literature.
The book of the day is The Feynman Processor : Quantum Entanglement and the Computing Revolution (Helix Books Series), and I'm going to give both a description of the book and a quick intro to quantum computing and my thoughts on the current state of the industry.
The book was published in '98, so it is missing some recent advancements in the field, but other than that is is pretty complete (up to at least '97). The merits of the book are that there is a great focus on just those principles needed in quantum computing, namely uncertainty, superposition, and entanglement. In fact the book starts out in examining information theory and the ability of the universe to create an endless or infinite stream of data (measure the same particle twice and you will get two separate results, hence an infinite amount of data). I think this is interesting to computer guy's because a lot of the discussion hits home. As the author moves into superposition and entanglement, he starts to re-use the same examples over and over again, in fact he does this throughout most of the book. For me, I get bored with the same examples, and since I understand the underlying concepts he is trying to explain I find myself wading through a lot of text trying to find any hidden meanings. For those not familiar with the principles I think the re-use of scenarios to explain each of the principles in turn is probably a good thing and creates a kind of thought parallel.
The most interesting chapter is on quantum software. He actually has a great examination of Shor's algorithm (better than anything I've found in the online literature), and about the various types of gates that will be used in quantum computing. There is a short examination of reversible logic structures and a direct parallel to the logic structures currently used in silicon based chips. He even explains the extreme heat produced by Intel chips in terms of Landauer principle. Can anyone honestly say they've ever given the mechanical reason for a processor being too hot when asked? Can anyone even say they knew the theoretical reason behind heat dissipation, and that a more complex chip could actually run cooler if it were made using reversible structures? Following quantum software he details three current approaches for creating a quantum computer (as of '98) that are promising. There are more now, but the ones he details are interesting enough. In fact it appears IBM did wind up creating a very small quantum computer using some of the methods detailed.
Based on all this, how in the hell do you structure a quantum computer? Well, very carefully. The power evident in the computer is that it is able to compute the possibilities over all inputs (a superposition) simultaneously into a set of all outputs (another superposition). Now, you aren't allowed to measure all of the states, only one of them. So while the quantum computer just computed everything, it'll only give you one thing (at least as far as we currently understand). The development of ingenious methods (entanglement) and formatting your algorithm (Shor's factoring algorithm) properly can all lead to some great speed-ups. If you take the Shor algorithm as a case in point, you'd have to compute a huge exponential expansion in order to factor the number using the same methods on a classic computer, while the quantum computer would do everything in a single operation. What they don't tell you in the beginning is that the it's frigin hard to entangle a bunch of qubit's together and current attempts only yield a couple to ten, when we actually need thousands of them for proper use. They also don't tell you that any outside interference into the process will collapse the superposition and ruin your results. They'll eventually work through the issues, they always do. The end result now is that there only exists a very small set of what you would call quantum algorithms. Factoring is obviously huge, as is searching over unordered data, and there are huge changes to the possibilities in cryptography. Beyond that not much else has been found.
Well, if you are interested in all those big words and really want to blow your mind imagining a miracle computer that computes all the answers you want, but only gives you one, and randomly at that, then read the book. As mentioned it isn't the best book I've ever read, but it does a great job explaining the basics of quantum computing.