Flower recalls the worst of that experience. Loose controls and a poor sense of depth. Entirely new movement aspects to consider like my turning radius. Homogeneous level designs with invisible walls and few landmarks. Undifferentiated paths and points of interest. Am I Mario or Lakitu? Am I the petals or the wind? Why can I turn around here but not there? Why can't I just slow down or turn the camera to get the angle I want to find this item or make that jump?
Despite that association Flower is not painful to play, and I want to identify why. Rewarding the players for exploration instead of punishing them for stagnation. The abstracted subject matter of the game. The pacing and integration of the themes - indeed, having a theme at all. Environments that are peaceful rather than frustrating to get lost in.
When someone asked me to briefly explain Flower's gameplay (ugh) I said "it's a game about making things better." I was referring to the game's environments at the time but after saying it I realized it went further than that. Traditional wisdom says polishing a game involves steps like "tighten up the controls, add more normal maps, and draw something really huge and evil-looking." Certainly that's one (expensive and difficult) development path. Flower shows us at least one way to step back, relax, reevaluate, and try to make it better.