At Chess Publishing Forum, IM John Cox has suggested a new 'shortest proof game' task: What's the shortest game ending with '0-0, mate' or '0-0-0, mate'.
So far the best tries are:
1.d4 e6 2.Qd3 Ke7 3.Bg5+ Kd6 4.Na3! Kd5 5.Qf5+ Kxd4 6.0-0-0, mate
1.f4 f6 2.Nh3 Kf7 3.e3 Kg6 4.f5+ Kxf5 5.Bc4 g6 6.d3 e5 7.0-0, mate.
Both achievements are quite good but I wouldn't be surprised if they can be improved upon.
One would think that short castling must take at least four moves: one knight move, one pawn move to open for the bishop, one bishop move and castling but that isn't necessarily so because Black can capture pieces too.
The same of course goes for long castling.
For those of you still toying with symmetrical mates from my blog entry of April 5th, here are some short games I found:
Knight: 1.e4 e5 2.Nc3 Nc6 3.Nge2 Nge7 4.g3 g6 5.Nd5 Nd4 (Dia)
Rook: 1.Nf3 Nf6 2.Ng5 Ng4 3.Nxh7 Nxh2 4.Nxf8 Nxf1 5.Ne6 Ne3 (Dia)
Bishop: 1.b3 b6 2.Bb2 Bb7 3.f4 f5 4.e3 e6 5.Be2 Be7 6.Bxg7 Bxg2 (Dia)
Pawn: 1.g4 g5 2.f4 f5 3.gxf5 gxf4 4.Nf3 Nf6 5.Ne5 Ne4 6.f6 f3 (Dia)
King: 1.f3 f6 2.Kf2 Kf7 3.Kg3 Kg6 4.Kh3 Kh6 5.e3 e6 6.Bd3 Bd6 7.Bg6 Bg3 8. hxg3 hxg6 (Dia)
Addendum July 12th
There is a quite readable article on Shortest Proof Games at Chessville. It appears to be the first in a series.