I'd say the same thing I always say.  Write the code, put it on a testnet, let people actually use it and prove beyond all reasonable doubt that it does what you claim.  Until then, no one is seriously going to support a change of algorithm to something unknown and totally untested in the wild.