What are the needs for the first version of the interaction vocab ?

[DrLeturc]

Hi all,

I'm opening this issue so as to get all the needs to release the first version for the interaction vocab.

Here I see the following points (please add the missing points or correct these points) :

• recognize the type of architecture (e.g. fipa) to interact with other agents / devices ;
• recognize the platform e.g. get the platform description to know the different services and how to use them, and the different devices an hyperagent can use e.g., their device profiles and affordances to interact with such devices ;
• recognize the different agents on a platform and the agent profiles, and affordances to interact with agents e.g. the message transport service (e.g. HTTPMTS / IIOPMTS) ;
• recognize the topology of the workspace, the workspace profiles and sub-workspaces to interact with the different artifacts in the (for physical?) workspace.

What I don't consider here :

• recognize the organization structure of hMASystem
• recognize the norms how to interact with other agents under the scope of a norm (we decided to consider this point under the scope of the organization structure)

But the organization part seems also an important part for the interaction vocab since it statutes also e.g. which agents can I interact with given the organization, etc. Can you clarify this point ?

[danaivach]

Hi @DrLeturc,

Indeed, establishing bridges from the Interaction Vocabulary to the regulation of interactions is in scope and should be further discussed. However, since there are yet no released versions of either the Interaction Vocabulary nor the Regulation Vocabulary, I propose that we don't yet establish explicit dependencies between the two since parts will be moving until release date.

E.g., recognizing the organization structure of an hMASSystem upon discovering how to interact with an agent in an organization could be addressed through a property like hmas-reg:isMemberOf or a property that relates an agent with the role that the agent enacts. Still, I would avoid creating scenarios for the Interaction Vocabulary that use such terms, until the terms can be found in a published ontology.

Similarly, specifying how to interact under the scope of a norm requires linking pins that should be discussed with people responsible for the Regulation Vocabulary once we have a first version of the Vocabulary.

Not establishing explicit linking pins between the two first versions of the vocabularies does not hinder the integrity of either since interaction can exist without being regulated, and MAS can be organized without direct agent-to-agent interaction.

CC: @gnardin and @andreiciortea in case you want to intervene.

[DrLeturc]

After the plenary, we decided to accept the following terms :

• Signifier ++ (introduction of skos:concept for Affordance )
• Behavior Execution
• Behavior Specification
• Action Execution
• Action Specification
• Input
• Ability
• Context
• hasForm
• expectsInput
• isExposedIn
• signifies
• hasAbility
• isEvaluableBy
• hasAbility
• expects
• [recommendsContext?]

[danaivach]

This is the agreed-upon list based on my notes:

Core Vocabulary

• Signifier ++ (introduction of skos:concept for Affordance )
• is Exposed in

Interaction Vocabulary

• Behavior Execution
• Behavior Specification
• Action Execution
• Action Specification
• Input
• signifies
• has Form
• expects Input
• has Schema

[DrLeturc]

5-10min to talk about this point during the plenary ? Can be quickly closed ?

[danaivach]

Current suggested list of terms:

Behavior Execution

Action Execution

Input

has input

Signifies

[DrLeturc]

We can close it since it was just a checklist. =)