I'm generally a fan of implementing spur of the moment ideas immediately while they're fresh (with the "iterate but move forward" caveat) as it keeps the energy and inspiration going, so I just implement them without even writing them down, or they go on the short list. If they're too big or I'm not at that stage yet I either break them down into smaller tasks or they go on the long list.
I don't believe in overly detailed prescribed design docs. I prefer to find the game as I'm building it and those lists are just there to give some direction, figure out the next step and to not forget any good ideas I have along the way. For my fish simulation game, Guppy, the long list was never more than a few pages of notes.