I really wanted to like this book, especially after all the glowing reviews it received. But the book has fundamental flaws in structure, topics, and approach:
- As someone else mentions, the book is heavily, albeit not exclusively, skewed towards Unix applications. Given that only 3% of desktops currently run Unix (6% if you're generous and include MacOS X), this immediately raises an accessibility barrier for those unfamiliar with the intricacies of Unix, and an unfamiliarity with Perl, Python, and shells. There are many Windows mapping applications that do much of what is described in this book far more easily and accessibly than the Unix-oriented solutions presented.
- Critical topics are defined incompletely and haphazardly. For example, while shapefiles are mentioned in several places early on, the first even barely-adequate definition comes several hundred pages into the text. Datums are covered very poorly, and given their importance in real-life applications, that's simply not acceptable.
- A significant fraction of the book is handed over to a discussion of different kinds of projections, and programs that will display them. While interesting, the fact is that most of the useful information about projections can be gotten from the text, or from a website that discusses projections; the "hacks" are essentially superfluous. I'm also disturbed by the amount of space given to projections vs. datums. In my experience, more people run into mapping issues with datums (e.g. using GPS units in WGS84, and wondering why they don't match up with topo maps in NAD27) than they do with projections.
- A huge chunk of the book is taken up GRASS. GRASS is very powerful, and as much fun to learn and use as sticking red-hot knitting needles into your eyes. Most of the problems that GRASS is used to solve in the book could far more easily be solved either with other freeware programs, or with Manifold GIS if you've got the money. For that matter, other difficult-to-use programs are employed to solve tasks that could be done more easily with other programs (e.g. SPLAT! is cited instead of the superior MicroDEM or RadioMobile for broadcast coverage; POV-Ray for 3-D models instead of 3DEM/Landserf/MicroDEM/Wissenbach3D; GRASS for map texturing instead of 3DEM or MicroDEM; PERL for spatial data analysis instead of GeoDA/STARS/PASSAGE; and so on). You could spend huge amounts of time becoming a "GRASS Ninja", as the book suggests, or you could spend less time and get more done using other software that's far easier to use.
- Sloppy terminology. For example, "Georeferencing" is used to refer to both linking digital photos to the location they were taken (should be "geotagging"), as well as its normal use in cartography, associating geographic coordinates with raster or vector features. And a GeoTIFF is not a Tiff with a world file, it's a Tiff with the georeferencing information embedded in it.
- A book of "hacks" should contain complete, self-contained hacks, and any number of hacks in this book don't fulfill this requirement. For example, the hacks on WMS/WFS, PostGIS, and building your own car computer/navigation system are either incomplete or hopelessly inadequate in giving you the information you need to implement them.
- Many of the hacks have been superseded by recent developments, like Google Maps/Earth, and the APIs from Yahoo and MSNLocal. This clearly isn't the authors' fault, but it does lessen the overall value of the book.
I could go on, but you get the idea. I have no idea of who the authors and publisher believe is the intended audience for the book, but unless you've got both a strong cartographic background and a strong Unix background, most of these hacks will be beyond you. And if you are strong in cartography and Unix, few of these hacks are even worth bothering with, since you'll most likely either know them, or know a better way to get them done. There are a few nuggets of useful information, primarily in the GPS section, but for most people, this book simply isn't worth the money.
Addendum (7/3/2011): The book is also now hopelessly out of date.