In the ACBL Bulletin, a problem hand was given. Something like xxxx-x-xx-AKJxxx after a 1H opening from partner. A large number of expert votes came in for a forcing 1NT, because focusing clubs even at the cost of burying spades seemed right.
If we assume this to be reasonable theory, an exception to the general Walsh thinking, then why not expand this further? It seems that the "impossible" 2S is not so impossible. Why not, in theory, 1H-P-1NT-P-2H-P-2S with 4-0-3-6 pattern? Focus the clubs, but mention the spades. Could not partner have 4-6-3-0?
If you take this out, then any 4-6 holding could be handled this way, and perhaps even 4-1-3-5 (perhaps passing 2D but converting 2H to 2S).
I am not sure where this thinking leads me, but the thinking is nonetheless suggested. Namely, there is nothing "impossible" about the "impossible" 2S.