At a local gathering of friends the other day, the classic puzzle book Maze, by Christopher Manson, came up in conversation. Many, myself included, recalled encountering it not too long after its original 1985 publication date. At the time, we all found it a fascinating artifact, though a completely inscrutable puzzle.
The book is still for sale, as it turns out, and there's also an online, hypertext version of the book you can wander through freely. (I note that the website appears to reside among the archives of an early electronic-publishing venture, and has remained unmodified since the mid-1990s. Sadly, the scanned illustrations are formatted to fit the relatively dinky computer displays of that era, resulting in much of the fine detail getting lost. I suppose I should encourage you to go buy the book, if you find them sufficiently intriguing.)
I should correct myself and call the book semi-scrutable, at least. It represents a labyrinth of connected chambers, you see, where each page features a haunting and evocative illustration of one room, trimmed with a short bit of text where the book's mysterious narrator leads a group of squabbling explorers through. The first part of the book's puzzle, then, is simply to find a path that takes you from the entrance to the maze's center and back in 16 steps. The harder part involves teasing the text of a riddle out of all the depicted stuff that lay along this route. And this is where most mortals get stopped, finding themselves with a pile of stuff and no clues.
After I returned home that evening, it occurred to me that I probably hadn't thought much about Maze since the ascent of Wikipedia, and surely it spelled out the solution. Why, yes. And what a solution! It's amusing I can look at this more than 20 years after the book's original publication and tell you why this would get razzed by any of the hardcore puzzle people I know today.
Granted, it was supposed to be very hard, because there was a cash-money prize for the first correct response. But the Wikipedia article implies that they overdid it, since the publishers extended the deadline at least once, and it's unclear if any claim was ever made. And no wonder, really; the solution demands you selectively perform wordplay on picture and text elements along the path, but gives you no clues as to which elements are important, and what should be done with them.
For example (and I'm about to get a little spoilery here) on this page, it happens that you're supposed to get a word by taking two picture elements and anagramming them together. But for all you know, maybe you combine the A with BELL and perform a sound-alike wordplay to get ABLE. Or perhaps the word is simply BELL, after all. Or a dozen other things suggested by the image. They all seem equally right - which is to say, none especially so.
Carry this feeling over the path's 16 pages, and I assert you've got an utterly unsolvable combinatorial explosion. I would be quite interested to learn of integral clues I'm overlooking, though, or to hear about someone who solved the book without any hints! Until then, I must conclude that for all the book's beauty - and it is quite a lovely thing to flip through - as a puzzle, it would get booed off the stage at the MIT mystery hunt.
More important than its puzzle, however, is the book's legacy. Without a doubt, the book left a lasting inspiration to many, stoking a hunger to try solving more baroque and beautiful puzzles, even if that means having to create them first. You can see echoes of Maze in art-heavy digital adventures such as Myst. In fact, the stimulus for this group recollection among friends was a new puzzle designed with Maze in mind, by Gameshelf pal Andrew Plotkin. I have it on good authority that it was cracked by dedicated solvers within a day.