I went deeper into Dusk’s rolling finality rules, and one detail changed how I think about “finality.”
A block is not simply final the moment it gets a successful attestation. Dusk distinguishes between accepted, attested, confirmed, and final states. An accepted block can still be replaced by a lower iteration block, while an attested block cannot be replaced by one.
The interesting part is how later blocks strengthen confidence. An accepted block becomes confirmed only after 2×n consecutive attested or confirmed blocks, where n represents previous non-attested iterations. Finality then depends on the parent already being final.
So the deeper question for @Dusk_Foundation and $DUSK is not simply “How fast is finality?”
It is: how should applications price risk while a block moves through these intermediate states?
For financial infrastructure, that distinction could matter more than a headline finality number.
How would you design an application around Dusk’s accepted → confirmed → final progression?
#dusk
A block is not simply final the moment it gets a successful attestation. Dusk distinguishes between accepted, attested, confirmed, and final states. An accepted block can still be replaced by a lower iteration block, while an attested block cannot be replaced by one.
The interesting part is how later blocks strengthen confidence. An accepted block becomes confirmed only after 2×n consecutive attested or confirmed blocks, where n represents previous non-attested iterations. Finality then depends on the parent already being final.
So the deeper question for @Dusk_Foundation and $DUSK is not simply “How fast is finality?”
It is: how should applications price risk while a block moves through these intermediate states?
For financial infrastructure, that distinction could matter more than a headline finality number.
How would you design an application around Dusk’s accepted → confirmed → final progression?
#dusk