also, the writing is quite terse. the lucidity is not really the strong point here. and sometimes they use symbols in proofs of theorems which were defined pages before. so, you need to be a bit attentive while reading. (someone might say, going back and revising the previous pages never hurt anyway).

despite, these small shortfalls, I have given this a four star because the book has a breadth as well as depth. lots of algorithms are presented and every presentation is complete. and also, this is perhaps the only good reference book on this topic, which makes it indespensible. after reading the book you will certainly be able to dive into the literature (papers on randomized and approximation algo), which is really what you want to explore eventually.