I've stopped using paper books for puzzles and now use only software. And then of course you always have a possibility to check an alternative solutions with an engine or even to play it out against an engine.
If your alternative solution of a tactical puzzle is inferior then with the help of an engine you can quickly understand why. With strategy puzzles it's much more difficult - sometimes I really struggle to understand why one variation is better than another. Ideally you should ask a strong player but not everybody has this possibility - I don't.
And there is yet another type of puzzles - endgame puzzles. Almost every book/database endgame position is a forced win/draw for one side. Playing for this side against an engine and trying to achieve the desired result is the best form of endgame training for me.
After playing with the puzzles in IM Silman's latest blogs, I realized that there are two types of chess puzzle (at least, for me): tactics/combination and strategy/continuation. I know that for tactics puzzles, simply do them thousands and thousands of times to improve. But! Besides mate in n moves, tactics puzzles often seems to have different solutions, especially if they come from a tournament games. So, should I check with engine for more variations or just solve it and move on?
For strategy/continuation puzzles, if the solution provided is considered the best, does it mean that my own variations are weak? I mean, if I am sharp enough to understand the purposes behind each move, I will probably be able to make it. I do believe that the moves reflect my styles/experience. So should I memorize the solution even if it does not fit my style or try to learn them and try to fix them to match my style?
https://www.chess.com/article/view/do-you-really-understand-positional-chess