There's a couple things rolling around the blogs and forums that I could write about, even want to write about, but I don't have the energy to tackle the whole subject. Stuff like character abilities and the sharp division of opinions on megadungeons. I think instead I'll just focus on one tiny subtopic of the megadungeon debate: the empty room.
Apparently, some people have problems with empty rooms. I think the problems come from two sources: GMs and players. On the GM side, I think some people are interpreting the admonition to only place monsters in one-third of the available rooms and treasures in one-sixth of the rooms without monsters a little too strictly. "No monster, no treasure" does not mean "empty", and in fact Underworld & Wilderness Adventures notably does not use the phrase "empty room", only the phrase "unoccupied room". There should be stuff in most rooms, preferably common, mundane stuff, with an obvious use, but not immediately useful, so that players can say "oh, this is the room where the goblins stored their pots and pans" and move on... until a later problem causes a player to say "hey, we should go back to that one room and grab some pots to solve this problem!"
But on the player side, I think too many players try an "exhaustive search" method of exploration. Move to Room 1, search thoroughly, fiddle with everything, and once you've done everything you can possibly do in that room, move on to Room 2 and repeat. That kind of approach is not only exhaustive, but exhausting. And if more than half the rooms are empty, the players are just going to feel cheated.
A better approach is to move to Room 1, make a quick assessment of the room (Occupied? Obvious container? Safe, or suspicious?) and move on quickly, basically doing a quick scan of part of the area before going back and doing a thorough search of anything. In essence, you want to secure an area before deciding which "empty" rooms ought to be thoroughly searched. Actual player mapping will help, here, because the players can skip searching for secret doors unless there's an area that looks promising: a gap in the map, a dead-end passage, four nearly-identical rooms, one of which is missing an exit... that kind of thing. The player map will also help define escape routes and defensible safe havens.