Glossary

---

####altruistic:#### we call a participant in some protocol or application "altruistic" if they are willing to perform an action or actions that benefit a protocol or its participants despite that action or actions incurring a cost to themselves. While protocols should not generally rely on altruistic participants, this term is useful when describing how a protocol might take advantage of the presence of one or more altruistic participants. The altruism of a participant will typically be bounded by some maximum of costs they are willing to incur, in which case we may refer to them as "boundedly altruistic".

---

####compliant:#### we call a participant in some protocol or application "compliant" if they do not deviate from the behaviour designated as "correct" in the protocol or modify the application away from its original specification.

---

####denial of service:#### also referred to by the abbreviation "DoS", this type of attack refers to a flooding of actions or requests which prevents some action from being performed, or results in excessive costs for or resource consumption by the affected party.

---

####fair:#### we say that a protocol or interaction is "fair" if there is no disadvantage to behaving "correctly" (however the protocol defines that). In other words, there is no "defecting strategy" which a malicious adversary can employ to depart from "correct behaviour" to the disadvantage of a compliant entity.

---

####local maximiser:#### a theoretical participant in a protocol or application who will never choose option "b" over option "a" if option "b" would result in a lower total gain (or higher total expenditure) of the economic units (and exchange rates, for example between computing cycles and assets) explicitly specified by the protocol description. We strongly prefer this term to the frequently used term "economically rational actor", because we want to explicitly acknowledge the fact that incentives or preferences not described in the protocol itself can always change the local behaviours of "rational" actors.

---

####malicious adversary:#### a theoretical participant in a protocol or application who will choose option or strategy "a" over option or strategy "b" if it increases the costs/harm or reduces the gains/benefit of another actor (usually a specifically designated one). Like the altruistic participant, an adversary's malice will typically be bounded by some maximum of costs the adversary is willing to incur, and such an adversary may be referred to as "boundedly malicious".

---

####preplay attack:#### when a counterparty in a state channel tries to publish an outdated agreement to the on-chain adjudication mechanism

---

####trust:#### trust is the ability for another entity to disadvantage you. If "A" trusts "B", then this means that B can perform an action without A's consent which would cause A harm or loss.

---

####trust free / trustless:#### a protocol or application is considered trust free if all other participants can be replaced by malicious adversaries without increasing the risk of loss or harm to a compliant participant.

---