Struct ConsensusSection 

pub struct ConsensusSection {
    pub self_weight: f64,
    pub susceptibility: f64,
    pub tolerance: f64,
    pub max_iterations: usize,
    pub susceptibility_of: BTreeMap<String, f64>,
    pub trust: BTreeMap<String, BTreeMap<String, f64>>,
}

Who listens to whom, and how hard the consensus iteration tries.

The trust rows are the influence graph DeGroot averages over. Each row names the agents one agent listens to, in whatever units the author finds natural: only the ratios matter, because crate::consensus normalises the row. An agent with no row listens to itself with self_weight and splits the rest equally over the others, which makes an unconfigured tracker report the tally as a fraction rather than something surprising.

[consensus]
self_weight = 0.5
susceptibility = 0.8

[consensus.trust]
reviewer = { maintainer = 3.0, worker = 1.0 }
worker = { maintainer = 1.0 }

[consensus.susceptibility_of]
maintainer = 0.2

Fields

self_weight: f64

Weight an agent puts on its own opinion when its row does not name it.

susceptibility: f64

How far an agent moves off the ballot it cast, in [0, 1].

One is DeGroot: an agent keeps nothing of its own starting position and the group converges on a single number. Below one is Friedkin and Johnsen’s generalisation, where each agent stays partly anchored to the ballot it actually cast, and what the iteration settles on is a profile of persistent disagreement rather than one shared position.

One by default, so a tracker that configures nothing keeps the reduction to the tally. Below one is the honest setting where reviewers are not expected to abandon their own reading, and it also removes the periodic case: any anchor at all makes the iteration a contraction.

tolerance: f64

Largest disagreement that still counts as settled.

max_iterations: usize

Rounds to try before calling the trust graph periodic.

susceptibility_of: BTreeMap<String, f64>

Susceptibility for one named agent, where it differs from the default.

Friedkin and Johnsen’s susceptibility is a diagonal rather than one number: a maintainer who has read the code for years and a reviewer seeing it for the first time are not equally movable, and saying so is the difference between the model and an average. An agent named here uses this value; every other agent uses Self::susceptibility.

trust: BTreeMap<String, BTreeMap<String, f64>>

Trust rows, keyed by the identity that holds the opinion.

Trait Implementations

impl Clone for ConsensusSection

fn clone(&self) -> ConsensusSection

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for ConsensusSection

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for ConsensusSection

fn default() -> Self

Returns the “default value” for a type.

impl PartialEq for ConsensusSection

fn eq(&self, other: &ConsensusSection) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for ConsensusSection