However, in reality it appears that merely hitting the target is not sufficient. One must also match the puzzle author's layout! In the puzzle I was playing, I managed to get the ball placed in a catapult which should have been able to hit the target. (And this placement was a correct part of the solution.) But no setting of the catapult strength resulted in a win; the ball appeared to miss every time. (The puzzle was "use that used thing 2 time.")
The problem turned out to be that I had not placed a roller element (a conveyer belt) which appeared in the author's solution. This turned out to be unnecessary, because a catapult could send the ball entirely over the roller and still deliver the ball to the next step in the solution. The original author set the catapult at a lower strength setting which touched the roller.
Now, to be fair, the rules do state that all the "bricks" have to be placed on the grid:
* The puzzle is only complete when all the building bricks have been placed on the grid and the ball hits the target.
But I think this is contrary to the spirit of the game. If the problem is that I haven't used all the pieces, the game should indicate that, rather than pretending that I have missed the target. In fact, I think this requirement is somewhat silly; it may be possible to create puzzles which are merely guessing games rather than engineering problems. Solving the engineering puzzle using the given piece set or fewer should be sufficient. Anything else breaks the physics of the game.