Agent
caution
This page has been moved to the eSSIF-Lab Framework on Github.
Short Description
An Agent is an actor that is executing an action action on behalf of some party (which we call the principal of that agent). During the time interval in which the action is executed, the actor may execute other actions in other execution-contexts, on behalf of the same or another party. However, for the execution of a single action, the actor is an agent for precisely one principal. It is assumed that the principal provides its agents with the policies that provide the agents with the rules, working-instructions, preferences and other guidance that they need to comply with when exeucting the action.
The Parties, Actors and Actions pattern provides an overview of how this concept fits in with related concepts.
Purpose
The ability to distinguish between (non)agents is relevant in many situations, including:
- electronic communication: the agent
Criterion
Agent is a role name that generically refers to an actor that is executing some action on behalf of a party.
Examples
- A person that is 'doing its own things' acts as an Agent for himself.
- A person that does things for his employer acts as an Agent for that employer.
- An ambassador, when (s)he is 'in function', acts as an Agent for the country for which (s)he is ambassador.
- A person that fills in the tax return form for someone else acts as an Agent for this someone else.
- A company that makes cars may use robots that weld parts of a car together; these robots acts as Agents for that company.
- A (running) webserver that accepts product orders for a retailer acts as a (digital) Agent for that retailer.
- A wallet app that runs on a phone and that is exclusively used by a single person acts as a (digital) Agent for that person.