A Different "Jacoby 2NT" Use?

After some thought following the usual BBF discussions, it dawned upon me that perhaps a better use for the "Jacoby 2NT" call is possible and worth considering.

The typical limitation upon Jac2NT for those who use good cuebidding methods seems to be a fairly balanced hand, ideally with primes.  This makes some sense, except that the equally modern trend for a 2C 2/1 as "real clubs or fit" or "real clubs, fit, or balanced" seems to cater equally well to these hands.  If you have a fairly balanced hand with primes, starting with 2C seems to work perfectly fine.  So, why Jac2NT at all?

In thinking through this issue, I realized that there are certain hand types that end up being "problems" when 2/1 auctions develop.  The "problem" occurs when we have a reduced likelihood of trumps being set at the two-level, usually caused because either Responder bids a 2/1 in the suit immediately below Opener's major (such that Opener cannot possibly make another call below his major) or, when Opener starts 1H, Responder is short in spades, such that Opener rebidding 2S is likely. 

For example, suppose partner opens 1S and you have 3-5-3-2 pattern.  If you respond 2H, the auction goes ballistic and spades cannot be agreed at the two-level.  With that specific pattern, I might opt to bid 2C, myself, as then I can switch tactics if Opener rebids 2H but otherwise usually can agree spades at the two-level.

But, consider a 1H opening.  If Responder has 1-3-5-4, for example, the auction is not likely to end up with hearts agreed at the two-level, both because Responder probably should bid 2D, which makes it impossible, and because even if a 2C response is selected Opener likely rebids 2S. 

Cuebidding sequences that start at the three-level are much less defined.  The solution might be to force certain patterns into two-level cuebidding by having 2NT be a GF raise with 3+ support and a trouble pattern, rather than Jacoby 2NT.

For instance:

1H-P-2NT = Hearts agreed.  Responder has a heart fit (3+) with long diamonds and/or short spades.
1S-P-2NT = Spades agreed.  Responder has a spade fit (3+) with long hearts and/or short clubs.

Re-defining 2NT along these lines then calls for different rules for the continuing auction, obviously.  One might have Opener usually bid a relay 3C to unwind?  For example:

1H-P-2NT-P-
3C(asking)-P-?

3D = long diamonds, balanced (2353, 3352, 2452)
3H = long diamonds, short spade (1354, 1453, 1552, 0454, etc.)
3S = short spade, with clubs (same as 3H, but with club length)
3NT = short spade, 1444-ish
4C+ = I have not worked this out that far -- what do you want from me?

Some similar type of unwind could be used when the opening is in spades.  In either situation, however, Opener mjight be able to break the relay, probably to show some very specific type of equally difficult hand contextually.

I have not worked out all of the possible sequences, as this is simply a brainstorming.  But, from the experience of actual bidding, I know that (1) Jacoby 2NT as balanced with primes is not that important any more, but (2) some hand types for Responder cause predictable problems that might be averted with the 2NT call re-defined.  Hence, this new approach might be worth considering and developing.