"Photoshop LAB Color: The Canyon Conundrum and Other Adventures in the Most Powerful Colorspace" by Dan Margulis is definitely worth reading and is most entertaining. I don't agree that it is "the most revolutionary book on digital imaging ever written" and it is certainly not a comprehensive look at Photoshop CS2. If you want comprehensive, get the gold standard: "Real World Photoshop CS2.
But there are some interesting gems buried in here that might improve your digital output. To me the biggest discovery was the idea of using curves in LAB color space to increase color variability. We are not just talking saturation here, but rather color separation. His explanation for why we want this is that cameras lack the sense of simultaneous contrast common to most human beings. When we see a lot of similar colors in close proximity, we break them apart. He shows how we can do that in Photoshop. It makes a big difference on some images.
His writing style is quirky, intelligent and often funny. That helps for such complex ideas. I think this book should only by used by advanced readers.
He relishes taking positions that are against the mainstream. For example, he argues that for photographic images high-bit editing is worthless. Likewise, he does not illustrate the use of Adjustment Layers or Layer Masks which are part of the standard workflow for most advanced users.
What is so wonderful about all his ideas, which go against the current accepted wisdom, is that he makes you question what you are doing. He presents his arguments using concise logic. It made me frequently pause and question my workflow.
Bottom line: This book should be at the top of the reading stack for Advanced Photoshop users. Learning how to increase color variability alone makes it worthwhile.
Henry Domke
Henry Domke Fine Art