Thanks kylemeister, yes 7. cxd5 is the main line of course, Houska does give Be3 as the best alternative for White though.  In my games cxd5 is pretty rare actually, I guess non booked-up Whites tend to assume that's what Black wants.

After 8. cd Bxf3 9. Qxf3 I think for want of anything better I have to try to snaffle the pawn with Qxb2 10. Rc1 Nxd4, White's lead in development is indeed scary but I can't see anything crushing, after 11. Qd1 Nf5 12. Bb5+ Kd8 13. 0-0 Nxe3 14. fxe3 Qa3 Black can't castle but maybe we can endure some unpleasantness and try to hang on to the extra pawn.

I just played a game where I played 7...Qb6 instead of the book e6, thinking I was in some other line.  In the game we both made mistakes, but the variations looked quite interesting, so I thought I would post them here.

The game went: 
1. e4 c6 2. d4 d5 3. exd5 cxd5 4. c4 Nf6 5. Nc3 Nc6 6. Nf3 Bg4 7. Be3 Qb6!? 8. b3? e6 9. c5 Qa5?! (Qb4 better) 10. Bd2 Bxf3 11. gxf3 Nxd4? 12. Nxd5? (12. Nb5 [I thought somehow Nxb5 would take care of this, lol]) Qxc5 13. Nxf6+ gxf6 14.Rc1 Qe5+ 15. Be3? Bb4+ 0-1 

It seems though the correct response is 8. cxd5:

8. cxd5 Bxf3 (...Qxb2? Na4) 9. gxf3 Nb8 (...Nb4 10. Bb5), and now I looked at 10. Bb5+ and 10. Qe2:

a) 10. Bb5+ Nbd7 11. Qb3 a6 12. Ba4 Qxb3 13. Bxb3 b5 14. Rc1 Nb6 and here my (old) engine wants to return the pawn with 15. d6 exd6 16. Ne4, which looks OK for Black. 

b) 10. Qe2 Nbd7 11. Rc1, I then looked at: 

11. ...g6? 12. d6 Rc8 13. dxe7 Bxe7 making g6 pointless;

11. ...Rc8?! 12. Bh3 a6 13. 0-0 Rd8 14. Rfe1 Qd6 15. Bg5 h6 16. Bh4 and Black is in trouble ;

11. ...a6! 12. Qc4 g6 13. Na4 Qa5+ 14. Bd2 Qd8 15. Qb3 b5 16. Nc3 Bg7 17. Bf4 0-0.

Black sacs a pawn, but seems to have good compensation in White's ruined pawn structure, and the two bishops are not ideal with the doubled d-pawns.

I have the feeling that a 7th move deviation in a main line can't be too great, but can anyone see any big problems with it?  All comments welcome.