[Englisch] [Gebundene Ausgabe]

Johnson's book is about a real quantum leap. The classical physics of our silicon computers does not hold within the tiny spaces inside atoms. Single particles at that scale can _really_ be in two places at once, and similarly, a quantum bit of information (known as a qubit) can be set to 1 and 0 at the same time, known as a "superposition." Qubits could be set to perform almost instantaneous calculations of huge programs, and there is no part of physics that says such computing should be impossible. Indeed, on the smallest of scales, primitive quantum computing has already been accomplished. Qubits are temperamental, and current research has to be done at supercold temperatures without the possibility of disturbance. Still, there is enormous intellectual interest in the prospect of quantum computing. One researcher in the field said that he and his colleagues are "writing the software for a device that does not yet exist." If quantum computing works, for instance, we will have to rethink all our current encryption methods, which are based on the difficulty of factoring large numbers; quantum computers do such things with ease silicon never can.

You aren't going to understand quantum computers by reading this book; Johnson knows that he is trying to describe the undescribable, and he makes it clear that he is no physicist, just someone trying to understand what all the fuss is about. His book is lucid and his descriptions do not bog down in technicalities (at times he gleefully hurtles over them). The book is also brief, but has enough substance to give even those who know little about current computing some basic understanding of where quantum computers may take us. He has successfully conveyed the excitement these potential gadgets have sparked, and readers will be able to participate in the excitement themselves.

Johnson acknowledges as much when he quotes French physicists Serge Haroche and Jean-Michel Raimond as saying that the small scale "hands-on experiments" with a few qubits that are currently being done "are more likely to teach us about the processes that would ultimately make the undertaking fail" than to teach "us how to build a large quantum computer." (p. 169)

As I understand it, basically the idea behind quantum compters is that (somehow) individual quanta (atoms, photons, electrons) are able to be in a particular state or not to be in a particular state; that is, either the equivalent of yes or no, but also in an indeterminate state; that is, a state that would signal yes and no at the same time! Somehow (and I hope I am forgiven for not fully appreciating this)--somehow because of this fabled indeterminancy, quanta can be used to compute at a speed that is more than exponentially faster than digital computers.

Johnson spends some series ink in trying to show how the atoms can hold and crunch numbers as long as they are not disturbed; that is, not measured in any way (which would bring about the famous "collapse of the wave function"). In this manner a problem that would take a digital computer weeks or months to solve could be solved in a fraction of a second. Problems now actually impossible to solve in any reasonable length of time might become tractable after all. The traveling salesman problem which grows exponentially more complex with the addition of each city, might very well yield to a quantum computer since the computational ability of a quantum computer itself grows exponentially with the addition of more quanta.

Wow. One of the reasons there is real money going into trying to develop these seemingly magical machines is that at present all the cryptography used by the military and big corporations relies on the fact that digital machines, no matter how fast, are not able to factor the codes. However, a quantum computer could. Furthermore, as Johnson explains, a quantum computer could also develop cryptography that could not be decoded. So, whoever gets there first--assuming somebody can--will at the very least make a whole lot of money.

What I found more interesting than the hope for a quantum computer are some of the insights into the quantum word that Johnson provides incidentally. The biggest stunner for me was his assertion that quantum events can be used to generate random numbers. It may come as a surprise to many people but in the world of classical mechanics there is literally no such thing as a truly random number generator. But because radioactive nuclei decay on a random basis, they can, according to Johnson, be used to generate random numbers. He writes that numbers generated in such a manner are "undeniable random." (p. 91)

Apparently this conclusion is a consequence of quantum indeterminacy. In a way, it is a circular conclusion since if we could somehow predict the rate of radioactive decay we would violate indeterminacy. I say "circular" when perhaps I should say "as a matter of faith" because there is no way a stream of numbers derived from radioactive nuclei decay can be proven to be random. Indeed, no string of numbers can, by examination, be proven to be random. If QM is true--and it is massively established--then the numbers are random.

Perhaps this idea of randomness is similar to the notion of "nothing" in that it is only defined in a negative way, by which I mean random is the absence of order, and order is in the eye of the beholder. What seems random to human beings may be quite orderly from another point of view.

Some of the book is pure fantasy. His discussion of quantum banknotes in Chapter 9 is an example of something that is useful to think about because of the light it sheds on the nature of the quantum world, but any chance that we would actually use quantum banknotes (requiring temperatures near absolute zero!) approaches the null set. (p. 146)

Other parts of the book are largely tangential (but interesting nonetheless). For example Johnson's exploration in Chapter10 of "nondeterministic polynomial-time" problems, such as the above mentioned traveling salesman problem, the protein-folding problem and the software verification problem, is very interesting. I was not aware that such problems were linked, but according to Johnson if one is solved, the others would yield as well. The current thinking is that the only hope of solving such intractable problems is a large-scale quantum computer. (p. 164)

Johnson is hopeful that such a computer can be developed and bases his hope in part on recalling just how intractable the problems toward the development of the sort of computers we have today seemed in the 1940s in the days of the vacuum-tubed Eniac computer which filled an entire room and had only a small fraction of the computational ability of my desktop. (p. 140) However, whether history will repeat itself and the impediments be overcome remains to be seen. It's exciting to think that they will.