Data Discovery
(
)
Most of the resources to dynamic plans do not reside in the agent itself but exists as external services and resources available in the Web. These resources can be classified as: Data from public repositories, such as definitions of the workflows and their steps, special rules, learning material, special rules, data of general interest or libraries that provide further functionality. Other public data include adding notification or news, to enable proactive services (such as notifying a new subsidy that applies to the user). Personal data, which is controlled by the user and accessed by the agent. Examples of this data are the user profile, the user preferences, health condition, economic situation, legal issues etc. Also, evidence documents can be stored to prove that a specific criterion is met, being of special relevance to dynamically compose workflows. The Data Discovery component looks for necessary data to execute steps by means of following links, invoking APIs or queries in a database. The component is supported by the reasoner for inference and mappings.