#dusk $DUSK spent last night rereading the dusk whitepaper section 0n plonk and something clicked that i missed...
the first two times through.
the proof scheme doesnt just hide values, it sparates what gets proven from what gets revealed. public inputs go one way, private ones stay.
locked ands the verifier only learns whether the statement holds true.
three properties have to hold for thats to mean anything: completeness. soundness. zero.
knowledgeness. skip one and the whole privacy claim falls apart.
whats interesting is this isnt bolted onto the state transition func tion after the fact.
its native to rusk vm. every execution path can carry a
proof of correctness without exposing thecomputation itself.
i cant fully resolve how proof generation cost scales as contract logic gets more.
complex though. thats the part i keep circling back to.
curious if anyone alse has dug into the verifier key setup here??
#dusk @Dusk $DUSK
the first two times through.
the proof scheme doesnt just hide values, it sparates what gets proven from what gets revealed. public inputs go one way, private ones stay.
locked ands the verifier only learns whether the statement holds true.
three properties have to hold for thats to mean anything: completeness. soundness. zero.
knowledgeness. skip one and the whole privacy claim falls apart.
whats interesting is this isnt bolted onto the state transition func tion after the fact.
its native to rusk vm. every execution path can carry a
proof of correctness without exposing thecomputation itself.
i cant fully resolve how proof generation cost scales as contract logic gets more.
complex though. thats the part i keep circling back to.
curious if anyone alse has dug into the verifier key setup here??
#dusk @Dusk $DUSK