See
BIP 103. The exact number and equation definitely need more research though.
It's an interesting proposal but it doesn't address the "market inefficiency" problem.
Because the BIP focus on problem of cost of running full node.
I think this is where I and ranochigo have a disagrement, I don't think Bitcoin will ever scale as a medium of exchange because our credit cards are just too convenient (and we generally don't have to pay any conversion/transaction fee to make a purchase with a CC)
It's certainly hard when people are too comfortable with the convenience.
but I do believe that it can answer a particular need and the block size limit can be high enough to answer that need without being too high.
I doubt it, unless combined with another solution such as Lightning Network.