Lewis Carroll wrote about logic.
If you're really interested in the historical treatment of this, you might find the pdf at the following link useful:
... and Dover published "Mathematical Recreations of Lewis Carroll" for a while.
Using the ISBN-10 ID:
This book doesn't exactly read like a "page-turner" novel. It is somewhat interesting from a historical, pre-computer age perspective.
Carroll alludes to the practical application of logic in contracts and laws.
Every time I try to read a bit more in this book, it is a chore, and I find something more interesting to do after a short while.
Also consider what it takes to construct a logic puzzle so that:
1. It has one and only one solution, and this property can be tested.
2. It doesn't require guessing at any point.
3. It has the "right" level of difficulty: Challenging enough to keep the interest of the best solvers and easy enough to encourage the relative beginners.
4. It has a sufficiently entertaining translation of math equations to "story" clues.
5. It can be stored, delivered, solved by users, and checked by a computer program.
6. It lends itself to machine generation to avoid the huge expense of manually constructing each puzzle.
If a given puzzle is just math, is it possible to measure the difficulty of a puzzle without a user-provided score or without a calculation of user completion rates and times? I.e, can the difficulty be calculated when the puzzle is constructed?
There are some interesting articles on the web that discuss this question.