How would you move mount Fuji?

By William Poundstone.

This little book discusses the use of logic puzzles for hiring programmers, and the mystery that they are apparently one of the better tools in the box to do this successfully. Especially the easier ones, which allow you to weed out ineptitude. Using super-hard ones to select excellence is not such a great plan, however, because often you have to have the one key insight, and it is much to random who of the smart people gets it.

It is also a collection of fun logic puzzles with answers. He is certainly right with some of the general strategies to employ, for example: decide if it is a plain riddle where you have all you need, and can just work it out, or if it meant to be a dialogue where you explore to understand the actual problem. if a puzzle seems to ask for inordinate amounts of calculation, then it is clear that you do not have to calculate much at all, that you should rather look for a trick or shortcut. If there are several options, and you can not know which is right because you seem to miss information, don't be paralyzed, crunch them through and eliminate the wrong ones as they lead to contradictions. Long complicated questions often have simple answers and vice versa. "Logically" acting people in riddles are not really people, you have to just execute the logical inferences, without ascribing common sense or emotions to them.