Kaplan Inversion by Opener?

Has anyone ever thought of this?  A Kaplan Inversion after a minor opening and One Heart response?

I was thinking tonight about the sequence 1D-P-1H-P-1S as opposed to 1D-P-1H-P-1NT.  For many of us, 1NT rebids might be made with a four-card spade suit.  1D-P-1H-P-1S, then, is unbalanced. 

Assuming this, I wondered whether a Kaplan Inversion by Opener might open up some interesting opportunities.  Opener rebids 1NT with an unbalanced minimum, four spades and five diamonds.  With the hand with which he would normally bid 1NT, he bids 1S instead, forcing to at least 1NT.  How would this gain?

Well, assume that Responder has the hand with which he would pass 1NT.  In that event, he instead simply bids 1NT after the inversion, and we wrong-side the contract, probably.  That's bad.

But, a lot of good things could then happen if we add in alternative holdings for Opener.  See, if 1S is forcing, then Opener can use this inversion as a relay before a next move with some of the tweener strong-ish hands.  Assuming, again, the simple auction, you have:

1D-P-1H-P-
1S!-P-1NT-P-
?

What can Opener rebid?

Pass = would have bid 1NT (probably wrong-sided)
2C = 18-19 balanced with either three hearts or four spades or both,  CHECKBACK!  If Responder has five hearts, he bids 2D as a transfer suggestion (Opener can reject this).  If Responder has four spades, he transfers to spades.  With neither, he transfers to 2NT and then can bid again as appropriate.  This allows nice stops on occasion at the two-level, a rare pass of 2C as well, and averts the somewhat preemptive 2NT, which now shows 2H and 2-3S and hence allows more focus for Responder's 3-bids.
2D = Intermediate, with diamonds.  This is the normal 3D call (jump rebid), but we can get out lower if Responder has a bust, and Responder can even rebid hearts.  This allows the direct 1D-P-1H-P-3D to be forcing, avoiding manufactured jump shifts.
2H = Intermediate hearts raise (anywhere from would accept a game try by Responder to would make a game try himself).  This allows stops at the two-level more often, more descriptive game tries (Opener is not forced to bid 3H with any number of hands but now has space to show stiffs and the like), and a different meaning for a direct 1D-P-1H-P-3H call (maybe GF).
2S = Spades and diamonds, enough for a reverse but nor for a jump shift.
2NT = Solid diamonds, needs something more.

Of course, we would need to work out the kinks as to what happens when responder has values or takes other actions, but this seems capable of working out.

Extend this, then, to 1C-P-1H-P-1S as another inversion.  This allows, after the 1NT relay:

2C = Club intermediate
2D = 3145 with extras but not good enough to reverse
2H = Invite raise
2S = Spades and clubs reverse-not-Jump-Shift hand
2NT = Balanced with four spades (direct 2NT denies four spades)

This time, the inversion solves the 3145 problem hand as the focus of the new minor call.  I mean, the same exact approach is possible, but perhaps it is more useful to handle this difficult pattern rather than super-tweaking strong balanced hands.  Plus, 2D gets messy anyway when you start thinking through the loss of re-transfers.  A smaller tweak works fine.