Oh sure, we've all got the Gang-Of-Four Design Patterns books on the shelf, right up there next to Knuth. I'd yank down my dusty copy whenever I needed to look up what a fellow coder meant by Facade or Visitor. (Actually, the short description of the patterns on the inside front cover usually was enough to fake my way through the rest of the conversation.)

In contrast, I charged through Head First Design Patterns in all of about two days. It was my first exposure to the breezy diagram- and photo-laden Head First series. You could consider the non-text portions to be just so much tree-killing fluff, but I found them a pleasant respite from what is, at heart, a pretty dry subject.

There were more than a few times during my reading that I sat back, whistled, and said aloud, "so that's how that works." The book covers the most common patterns from GoF in an incremental order. I was disappointed that some patterns were lumped in the last "Leftover Patterns" chapter because I would've enjoyed the authors' take on them, particularly the Flyweight pattern, a personal fave.

Examples are illustrated using Java. That's definitely an improvement over the templated C++ in GoF, but it does illustrate a failing: the old-school object-oriented languages like C++ and Java needed patterns to solve common problems. The latest batch of OO/functional languages like Python and Ruby have little use for some patterns, and add new patterns all their own. For instance, what use is there for an iterator pattern in Ruby that uses closures to loop? Why bother with factory patterns in languages with first-order functions and class objects?

That opinion aside, patterns are still an everyday matter for the OO practitioner, and Head First Design Patterns is a superb introduction to them.

The GNU/Linux bash shell is a clunky marvel. Novices recoil at first contact; “how do I click my way out of this 70s-era greenscreen abomination?” Casual Linux admins--and I fall into this it-was-set-up-a-year-ago-and-still-works-fine-so-don’t-futz-with-it category--are reasonably adept at piping, redirection, and tab-completion. The full-bore Linux geekorati are only a couple of Emacs meta-ctrl-popbottle keystrokes from involuntary carpal retirement.

Learning the bash Shell is really intended for the second of these groups: the not-everyday Linux enthusiast. The experts will have already glommed every tip and trick, and, despite the disingenuous "Learning" in the title, the book’s too steep a road for folks still struggling to install their first RPM.

O’Reilly is known for its content-dense publications, and this book has a higher fact-per-unit-volume ratio than any other of that publisher’s titles that I could name. Like many a tech trade tome, its chapters should be read as you need them, not straight through from copyright to colophon ("Typeset with ITC Garamond, you say. Fascinating!"). I go ahead and store it right next to the server. A quick peek will tell me everything I need to know, and little I don’t care to learn, about test/[] switches, string substitution operators, special-case environment variables, file descriptor redirection, Emacs control commands, process substitution, and those darned umask settings (each an example of something I use often, yet the details of which I can never recall). And once you’ve got the book open, you’ll find it just leads you further and further down the bash rabbit hole.

In short, this is one of the best tech books I’ve ever encountered, for any OS, for any topic. If you’re of the aforementioned casual Linuxfolk, or transcending your way to the guru plane, you must add it to your library.