Why read this book now? Nowadays, when I want to make a crossword puzzle, I load up Crossword Compiler, click a button, and let the computer fill in my grid with answers that, y'know, cross. (OK, I'm exaggerating the ease. Since I work on puzzlehunts, I make puzzlehunt-ish crosswords, by which I mean "gimmick crosswords". To make the gimmick work, I probably need to do some prep. Like if the gimmick is something numeric such that LONDONENGLAND should be written in the grid as LOND1NGLAD and FORTWORTH as FOR2RTH, I have to start by constructing a word list so that Crossword Compiler understands that, for this puzzle, LOND1NGLAD is a valid answer.) This book has plenty of advice that I won't use. E.g., when fitting answers into a crossword puzzle by hand, favor answers that alternate vowel/consonant, since these tend to fit into crosswords better. That's a good rule of thumb when you're working by hand and using your brain; but when a computer can try all the possible "crossings" in a minute, it's less useful. And yet, this book's still useful today.
Crossword puzzle fans will point out a flaw in my reasoning above, and in so doing point out the first useful thing I learned from this book. Crossword puzzle fans will point out: The best crossword puzzles are hand-constructed. So the whole premise of saying that the book might not be useful since it concentrates on by-hand puzzle construction must be wrong. Well… yes and no. I make crosswords that will be used in puzzlehunts. People don't complete these crosswords. They fill in a third of the answers, someone sees F□□R■N□M□D■S□□P in the diagonal, figures out that's FOUR-NAMED SOAP, and folks set aside the grid, figure out DAYS O4 LIVES and… Darned few folks will go back, look at that grid, fill the rest in, and tut-tut over mediocre word choices. I can get away with some things that I couldn't if I were a professional puzzle constructor trying to sell stand-on-their-own-merits crosswords to Will Shortz. And this book told me something I didn't know about making crosswords by hand: even if I were to get good at it, it would still take a really long time to make one puzzle. The book doesn't really point this out explicitly, but you can't help but notice the word "hours" popping up. Hours for this stage of construction; hours for that stage.
With practice, crossword construction gets easier, but it doesn't get easy. Good to know. If you don't plan to make enough crosswords in the future to justify training up, need a high-quality grid, and you value your time, maybe you should hire a pro.
But even if you're just jockeying a computer, this book will help. It has advice on what to do if you get partway through constructing a puzzle and it's just not coming together. This happens on the computer, too; it's just faster. Sometimes you choose a grid, lay in your theme answers, press the "fill" button and… the computer gives up. Or it generates something that's bad even by my low standards. Maybe you can salvage your idea by tweaking the grid: when you laid your theme answers into that arbitrarily-chosen grid, did that mean the computer was going to have to choose a word ending in J? Maybe you should figure out a grid layout that puts that J at the start of a word instead. (Crossword Compiler does a great job of "trying all the words" to find those that will fit into a grid; it doesn't have a notion of "trying all the grid layouts" to find one that will work best with a set of theme answers.)
This book talks about the core of puzzle construction; even though I skimp on the details of exquisite puzzling, there was stuff in here I could use. And if you're a by-hand puzzle constructor (or if this book inspires you to become one), there's even more good advice.