## Keyboard shortcuts

Press `←` or `→` to navigate between chapters

Press `S` or `/` to search in the book

Press `?` to show this help

Press `Esc` to hide this help

- Auto
- Light
- Dark

# Algorand Specifications

The resynchronization is an auxiliary function used throughout the recovery steps.

A partial order relation is defined in the space of all observed _bundles_. We call this relation _freshness_.

A resynchronization attempt broadcasts the _freshest_ observed bundle (if any).

Priority-wise, bundles’ freshness is defined as follows:

- Bundles for a certcert step are fresher than all other bundles.
- Bundles from a later _period_ are fresher than bundles from an older period.
- Bundles for nextnext step are fresher than bundles for a softsoft step of the same period.
- Bundles for nextnext step for the ⊥⊥ _proposal-value_ are fresher than bundles for a nextnext step for some other value.

> For a formal definition of this property, refer to the ABFT [normative section](https://specs.algorand.co/abft/abft-player-state#special-values).

> Important
>
> **IMPLEMENTATION:**
>
> Bundle freshness [reference implementation](https://github.com/algorand/go-algorand/blob/55011f93fddb181c643f8e3f3d3391b62832e7cd/agreement/player.go#L518).

> In the reference implementation, a resynchronization attempt is handled by the `partitionPolicy` function, as the network is assumed to be in a “partitioned state” due to the temporary inability to reach consensus. In this case, the function is only invoked when the current step s≥3s≥3 or when the current period p≥3p≥3 (that is, the player has gone through _two full periods_ without reaching a consensus).