Good question, I guess it's more realistic to have no hint, and it can't be bad to simulate a real game situation. Ray Cheng's "Practical Chess Exercises" is such a book where each diagram only reads who is to play and you don't even know if it's a positional answer or a tactic.
But I think traditional puzzles are good too, and I guess it shows through their popularity over all these years. Yeah it gives hints that are unrealistic to actual game play, but it helps the student focus on a particular area such as mate in 1, mate in 2. Even by theme such as Reinfelds puzzle book breaking it down into pins and forks. It helps the player burn a certain type of pattern into their brain.
For example, a beginner going though Cheng's book may be impressed by the idea of a skewer, only to have to wait 100 pages to see an example of it again.
Does working on puzzles without clues e.g, mate in 1, white to move and draw, make your brain work harder ? Is it good for training and useful ?