As thorough Khalifman usually is I expect that even 6..a5 should be covered in his book!? --as this line would simply avoid his preferred variation...
Anyone knows?

So far we have been talking about: 1.d4 e6 2.c4 f5 3.g3 Nf6 4.Bg2 Be7 5.Nf3 o-o 6.o-o a5!?
What if white plays 6.Nc3 instead? Are there any reason to not play 6..a5!? here as well?
The only reason I could think of (now) is 7.Bg5!? (with the plan to Bxf6 and break up the centre with e4). Normally black has Nbd7 available. I have to look into this....

I don't own the Chalifman book, but I recall having read, that he advocated either 7.b4 or 7.Nc3 and 8.b4.
If you don't like it as Black, you can avoid it: 1.d4 e6 2.c4 f5 3.g3 Nf6 4.Bg2 Be7 5.Nf3 o-o 6.o-o a5 followed by 7...d6. White's only independent option 7.d5 does not look scary (Na6 or Bc5), now I have looked at it. This was pointed out by Richard Palliser above.

You have already answered your own question. 7.b4 and the related 7.Nc3 Qe8 8.b4 is in my opinion one of the two ways to play for an advantage for White.
Williams' work is certainly worth buying, as it urged and urges me to work on my pet defense. But the b4 variations are not the only insufficient treated IMO.
Fairly typical for Williams is, that against one he recommends e5 exchanging the e-pawn against the White b-pawn. Against the other he condemns it.
Having said this, I must stress, that Williams has included a lot of original ideas ans analysis. In the process of thinking them over, he has certainly improved my understanding. That is much more, than many other writers can say.

Yes, I saw that post by Richard Palliser and it did set me thinking, but so far I haven't seen 7.b4 in my games anyway, so I'm not sure how I'll handle it when it crops up.  I ought to mention that I am playing either Blitz or email/CC chess at low levels so the theory isn't as important for me - most of my opponents deviate well before move 7, actually! That's not to say that I won't take the theory seriously, at least up to a sensible point - I like to know what I'm doing - but I haven't experimented too much just yet.  Lots of work still to do...

Right.  Well at the moment I'm going with 7...Ne4 in my games, and so Williams seems to have the best coverage of this, which I'm using, and he's also given me the fundamentals of the opening in clear enough terms for me to get a handle on it.  But I think I will eventually end up playing 7...a5, and for that I'll probably supplement Williams with Pinski - I think that will be the right time for me.  I'm in agreement with Williams' view that 7...Ne4 seems to be the easiest line to learn, but I'm also starting to appreciate the point made by MNb in another thread about the problems associated with allowing one of your pieces to be exchanged off early in the game, when it may be more useful to hold onto it for use in an attack on White's king.  I think this opening lends itself to keeping the pieces on the board for as long as possible.  This is something that is perhaps worth my bearing in mind in all my games - when I used to play the QGD Tartakower I would often aim to exchange down into an endgame and hope to grind out a win with what was left on the board, until I read in John Watson's second 'Strategy' book that contemporary methods of playing the Tartakower often aim to do the opposite of that, i.e. keep the tension by keeping the pieces on, and aim for dynamism rather than an endgame.   I think that now that I have moved to the Dutch this is especially relevant - my tactical play seems to be improving a little, and I'm certainly having more fun with chess lately.

Thanks bravehoptoad.   

Useful info, and nice also to see that someone likes the Pinski book which mostly seems to loiter in the shadows behind Simon Williams' work on the Classical.  I think I'll get myself a copy once I'm done with my other sources.

Yes, I think I'd rather avoid the Pirc.  I've been looking at the French lines MNb mentioned, but I still think it might be more logical to just play 1...f5.  I had a look at the 2.Bg5 lines again last night and I think I could get used to them - but as TomR says, 2.Nc3 looks quite tricky.  In Williams book he recommends 2...d5, which kind of emulates a Stonewall, and so I had a quick look at this last night - it seems solid enough, but in McDonald's book he strongly advises against playing 2...d5 and prefers ...Nf6 instead, saying that after 2...d5 Black faces a thankless defensive task with few winning chances.  Williams, on the other hand, is in favour of 2...d5, saying that 2...Nf6 simply allows White an easy route to a small advantage, usually via the doubling of Black's pawns on f6.  Personally I prefer the look of 2...d5 at this stage, although I haven't finished working through the lines yet.  I wonder if anyone could tell me what Pinski's view on this is - his book on the Classical Dutch is the one that's missing from my library at the moment....?