Rating: 

I think the essential problem is bundling together awk and sed, two philosophically very different tools. The result seems to be a book that is bitty, has no clear narrative line, with arbitrarily chosen examples that seem to be hard to track down when you're trying to use the book for reference. What I mean by that, I think, is that although the detailed discussions are useful and instructive if the problem you're trying to solve is a close match to an example focussed in the book, it's not at all easy to derive general principles, or extrapolate from the rather narrow selection of cases under consideration.

The problems inherent in writing a book about awk (and gawk, and nawk...) are a difficult enough hurdle by themselves. The addition of sed was a language too far. And heaven's knows, sed -- an indispensible tool, but a formidably difficult one for beginners and cognoscenti alike -- surely deserves a book on its own.

The O'Reilly UNIX books set a very high standard. I'm sorry to say I don't think this one quite makes the grade.

el bid