Richard Dawkins' The God Delusion is an important book: There have been other books about atheism, but most are written in such a way to preach to the converted, or have been ignored. His book has succeeded where they have not, and for that, Dawkins deserves all respect.
Dawkins sets out his purpose in writing this as to make it clear for people stuck in religion that there is an alternative, and they have another option that is, indeed, an acceptable one. As someone who spent years tormented because I never knew that my growing discontent with the problems I kept finding in the weird branch of fundamentalist Christianity my mother espoused, and feeling I had noone to turn to (while never realising my father was a damn Freethinker, and I could have gone to him at any time. Sheesh, dad! You could have dropped a few hints, and not suddenly tell me when I'm 27 and have worked my way through it alone), I can only wish I had known about this book, or any other like it, as a child. (He wanted me to be able to make my own decisions, and so just made sure to teach me science and mathematics, and so on. I appreciate this, but, still...) This book fulfils its purpose as set out, and thus must be considered a total success.
However, it does mean that like "The Selfish Gene" and "The Blind Watchmaker", I'm going to be reading this book far too late in my education for it to really tell me much I don't know, so I'm going to end up much more critical than other supporters would be.
Chapter 1 deals mainly with the Einsteinian version of God - Spinoza's pantheism, or God as a metaphor for the universe and its laws. It's actually very well done - engaging, and also informative.
However, like the rest of the book, it's somewhat sloppy. References are given in three seperate forms: Footnotes, endnotes, and worked into the text; none of them would allow you easily locate a quote Dawkins used, unless the source is short, because Dawkins never gives page numbers. On page 16 (British hardcover edition), he begins to describe reactions to Einstein's public statement of his beliefs, gives a source for them, then parenthetically mentions that this source was his "main" source for all the quotes that came before. Does that mean some of them are not from that source? If so, where? Dawkins is silent.
I trust Dawkins, and am willing to believe he's gotten things mostly right, and don't feel the need to check he's accurately quoted Einstein. However, a standard rule of referencing is to be sure to cite anything that's especially difficult to believe. A section on the reaction to the Danish cartoons (page 25) claims that one sign at a protest read "Behead those who say Islam is a violent religion". This complete lack of irony is especially difficult to believe. Now, this is tangental to his main point, and three references do support major points in his description of the affair from Danish imams intentionally manufacturing protest and adding three additional images to the original set in order to provoke anger further. Sadly, I can easily believe the uncited, but plausible descriptions of violence that occured as a result of it. However, while getting upset about a few cartoons may not show intelligence, "Behead those who claim Islam is a violent religion" is ridiculously stupid, and therefore ought to have been cited.
Chapter 2 begins by setting out the existance of God as a minimal hypothesis, "There exists a super-human, upernatural intelligence who delibrately designed and created this universe, including us." He goes on to mention that individual religions add rather a lot of extra baggage to the hypothesis wihich makes it more and more improbable. He mentions polytheism, but explains that the arguments against it aren't significntly different than that of monotheism, and most of his readers are probably more familiar with monotheism. This is... to some extent fair enough. He also explains that the Trinity, angels, and saints are polytheism in all but name (an opinion I've long held myself). He deals at length with the deism and atheism of America's founding fathers, and their probable horror at religion taking over America, and it's great reading, and very convincing stuff. A summary of an incident from David Mills' book "Atheist Universe" on page 44 was very well chosen, but perhaps a bit too well chosen: it's far more exciting and readable than anything Dawkins writes himself, precisely because Dawkins has almost certainly never suffered significantly for his atheism, but Mills lives in America and has. Dawkins' detached style cannot compete with a good personal story.
Then he spends 9 pages on an unconvincing and, frankly, ridiculous discussion of "why the term agnostic is bad." He explains the source, T.H.Huxley, one of my favourite writers... then... redefines agnosticism as saying that the two possibilities are equally probable - that God's exostance and non-existance are both equally likely to be true.
This is poppycock. Unadulterated nonsense. Huxley was seperating himself from people who rejected religion outright on emotional grounds. He was defining agnostic as someone who analysed the evidence, and could find no evidence that God existed. Dawkins pulls his equal-probability claim out of thin air. This poorly-argued bit of claptrap should never have made the cut to reach the final book.
The second half of this section quotes half a dozen examples of undisprovable, but highly improbable things as supposed examples of why agnosticism (under his definition) with regards to questions of faith is foolish, and how the proper term is atheism. These range from Bertrand Russel's teapot orbiting between the Earth and Mars to invisible, intangible, and inaudible unicorns to the Flying Spaghetti Monster. Also "the world is rhombus shaped and borne through the cosmos in the pincers of two enormous green lobsters" - which is highly disprovable, even easily disprovable, and should not have been mentioned.
This being Dawkins, he then has to spend a section attacking Gould. The concept of "non-overlapping magesteria" from Gould's "Rocks of Ages" (a book I admire), was one of Gould's many attempts to put religion in its place by defining it as precisely equal to ethics, and keeping it out of science.
Dawkins, however, spends 7 pages completely misconstruing Gould's points. Again. Is some great public demonstration of his inability to follow the logic of anything Gould writes at least once per book a requirement of his book contract or something? To be fair, though, his points in this section are perfectly valid, just directed at something Gould, a fellow atheist, didn't believe in the first place - which Dawkins admits on page 57 "I simply do not believe that Gould could possiblyt have meant much of what he wrote in Rocks of Ages", but goes on to say he, as Dawkins claimed Huxley did in his setting out of the meaning of agnosticism, was "bending over backwards to be nice to an unworthy but powerful opponent". Dawkins best evidence for this is that Gould says that we cannot comment on the question of God's existance as scientists. Dawkins reply that we can, however, comment on its probability. However, the key phrase is "as scientists": science does not work by what seems most likely or ought to be true. It works by making hypotheses, collecting evidence, and testing them. Dawkins' makes several arguments against this view, but all are very poor, and get increasingly far from anything Gould ever claimed as they go on. They all boil down to either "some things could potentially be proven one way or the other if we somehow got evidence" or simply saying that all logical endeavours count as science.
This section is his most flagrant example of lack of references: "How many literalists have read enough of the Bible to know that the death penalty is prescribed for adultery, for gathering sticks on the sabbath, and for cheeking your parents." The relevant verses are not given, though it's implied they're in Deuteronomy or Leviticus.
I'll lend a hand. The verses in question are Leviticus 20:10, Numbers 15:32-35 (Guess the implication was wrong), and Deuteronomy 21:18-21. Always make it as easy for someone to confirm a surprising revelation as possible. Being undeniably shown that sort of thing is in the Bible is a good way to demonstrate against inerrancy. Making vague assertions that sort of thing is in there is not.
However, I'm happy to say the rest of the chapter returns to the engaging and highly readable mood with which it opened.