@Dusk_Foundation
I kept coming back to the word “confirmed” while tracing Dusk’s finality path.
The surprising part isn't the four states.
It's that the path to confirmation changes depending on what happened earlier in the round.
In the rolling-finality model, the first iteration starts with n = 0 previous non-attested iterations, so it gets the fast path.
Now let two iterations fail to produce the required attestation.
n = 2.
The rule becomes 2×n, meaning four consecutive blocks with the required attestations or confirmations are needed before the block being evaluated becomes confirmed.
What caught me is that this happens when the round is already behaving badly. Confirmation can actually require more evidence before it moves forward.
The round's history changes how much evidence the next block needs.
So confirmation isn't just about the block. It's partly about what the round did before it.
What I still can't tell from the paper is how often that extra depth appears under real network conditions.
$DUSK becomes more interesting to me if this adaptive finality stays predictable when the network gets messy.
#dusk
I kept coming back to the word “confirmed” while tracing Dusk’s finality path.
The surprising part isn't the four states.
It's that the path to confirmation changes depending on what happened earlier in the round.
In the rolling-finality model, the first iteration starts with n = 0 previous non-attested iterations, so it gets the fast path.
Now let two iterations fail to produce the required attestation.
n = 2.
The rule becomes 2×n, meaning four consecutive blocks with the required attestations or confirmations are needed before the block being evaluated becomes confirmed.
What caught me is that this happens when the round is already behaving badly. Confirmation can actually require more evidence before it moves forward.
The round's history changes how much evidence the next block needs.
So confirmation isn't just about the block. It's partly about what the round did before it.
What I still can't tell from the paper is how often that extra depth appears under real network conditions.
$DUSK becomes more interesting to me if this adaptive finality stays predictable when the network gets messy.
#dusk