#dusk $DUSK @Dusk There's a Dusk audit finding that makes the “verify by arithmetic” idea more interesting than it first appears.
The argument is simple: emission is fixed, rewards are proportional to stake, so anyone can reconstruct what a validator earned from public numbers. No dashboard. No trust. Just arithmetic.
But there's a layer underneath that assumption: are the numbers themselves correct?
Oak Security's review of Dusk's consensus layer and Rusk node library found validation logic issues in the layer responsible for determining whether state transitions are valid.
They were caught before mainnet and fixed, with all critical/major issues resolved before launch.
That's good. But it highlights an important distinction:
Transparency and correctness aren't the same guarantee.
A ledger can make its numbers completely visible. That doesn't automatically mean the logic producing those numbers is flawless.
The “verify by arithmetic” property is valuable precisely because the underlying consensus rules are correct. An audit helps establish that — but an audit finding also shows why that guarantee can't simply be assumed.
So where would you draw the trust line?
A second independent audit, years of clean production, or something else?
#dusk @Dusk $DUSK
The argument is simple: emission is fixed, rewards are proportional to stake, so anyone can reconstruct what a validator earned from public numbers. No dashboard. No trust. Just arithmetic.
But there's a layer underneath that assumption: are the numbers themselves correct?
Oak Security's review of Dusk's consensus layer and Rusk node library found validation logic issues in the layer responsible for determining whether state transitions are valid.
They were caught before mainnet and fixed, with all critical/major issues resolved before launch.
That's good. But it highlights an important distinction:
Transparency and correctness aren't the same guarantee.
A ledger can make its numbers completely visible. That doesn't automatically mean the logic producing those numbers is flawless.
The “verify by arithmetic” property is valuable precisely because the underlying consensus rules are correct. An audit helps establish that — but an audit finding also shows why that guarantee can't simply be assumed.
So where would you draw the trust line?
A second independent audit, years of clean production, or something else?
#dusk @Dusk $DUSK