1.e4 f5 : me and my oversized ego vs all of you
Good move. Although those who liked f4/f5 for you in some lines will be disappointed.
A pawn down I don't fancy going into an endgame. But that e6 pawn can't be left there too long.
The horns of a dilemma.
What will the engines have to say?

Sjeng is really amazing. It wants to play Qxe6. I can't imagine it will get support from any other engine for that move. Its alternative is Bxe6.

Just a brief explanation what does h6 cause?
In case it's a genuine question, which I doubt : suffering, starting with Bg6+.
At depth 25, Houdini 1.5 gives 9...0-0 at 0.81 with 9...Bxe6 trailing behind at 0.83 (although the lines do transpose).

Irontiger, I just quickly looked at the position and forgot about the other bishop. However, as a "human", I'd probably play O-O.

Not arguing with 0-0, just curious. If you guys were playing against a human at about your level, would you 0-0? White has several pieces aimed in that direction, and could soon add a rook to the party. My inclination would be to play Bxe6 & Nc6 (not nessesarily in that order) and 0-0-0.
I would play Bxe6 as well. 0-0 feels too passive. I guess 0-0 is OK afterwards. It's true many pieces are poised to attack, but we have many pieces ready to defend in the area. I also like the fact that our rook will seize the f-file. Of course it can backfire because the position is difficult to play, but not for the engines.
0-0-0 woud take too long, we might never be able to castle queenside or kingside at all.
MSC157, I did what you advised, and I tried to teach the engine the position, I only showed the engine 0-0, but it changed the evaluation from 0-0 to Bxe6 up to ply 28 when it got back to the original 0-0. Very useful advice. Next time I'll show it more positions.
Seems that 0-0 is the engines' choice.
I liked it last time so I guess I still do - although I had my eye on the e-file and the OP's queen and king invitingly lined up whereas now that file is very congested plus the OP is a move closer to castling himself.
Still it is a nice developing move and I agree that castling long is not an option and that getting the rook onto a half open file is a distinct plus.
I have played a number of games in which a decision to delay taking an exposed pawn comes back to haunt you later. But perhaps that won't be the case here.

I have played a number of games in which a decision to delay taking an exposed pawn comes back to haunt you later. But perhaps that won't be the case here.
The idea is to take on e6 right after 0-0.

@LongIslandMark: I think that the king is more safe at the king side than at the queen side, because white will not likely start playing his king side pawns. He would definitely bring in his queen side pawns, if we would play 0-0-0. I would therefor play 0-0 and Bxe6 in this position.
But I would have never played 5... c4. My preference would be 5... e6.
0-0 doesn't look good because Black's king might get caught before his pieces can rally to his defense. It has the right idea (protect the king), the king-side is weak for black, so I think we should play Qe7 first and see where we go from there.