Quick thought: When considering new mechanics to add to a puzzle game, consider how they can be useful AND detrimental to the player’s goals. Ie. ask yourself, does the mechanic have both pros and cons? If so, design puzzles that demonstrate both sides of it! If not, could you modify it in some way to make it two-sided? If they only have cons, the player may feel annoyed. If they only have pros, that’s probably fine, but less cool 🙂
Example of a con-only mechanic: The concrete un-portal-able walls in Portal 1/2. Portal 2 used them way more, and while it makes for some neat puzzles, I couldn’t help but feel that they were a bit cheap. Especially considering that Portal 1 used them in relatively fewer situations. Could they have been useful somehow? What if there were “enemies” of some sort that shot portals, and maybe in one puzzle you had to prevent them from portaling? That’s not really a great idea, so perhaps this is a lost cause. I’m sure the Valve folks pondered these questions as well!
Example of a pro and con mechanic: The magic items in Braid that were unaffected by your rewind. They were useful, such as the first key-in-pit puzzle, but they were also detrimental when applied to enemies and certain platforms. Same with the goo’s in Portal 2: Clearly they were useful in a lot of situations, but in others they acted as an obstacle. I’m pretty sure there was at least one puzzle where the bouncy-goo was in your way and you had to find some way around it.
In my current game, I was about to add lava – ie. parts of the level that will kill the player upon touch. But then I asked myself..is the death necessary? If I removed the death, and just made it rock that is unaffected by the reflection mechanic, it can still function as an obstacle by blocking the player’s path. But then it can also help the player as a platform! So I’m gonna implement it without death, give it collision, and try to design puzzles that also highlight its positive uses.