Should we explicit the set of personas (here agents) like Polifonia ?

[DrLeturc]

Polifonia considers explicitly a set of personas with roles. These personas create stories (i.e. motivating scenarios). It may be relevant to explicit research-based description of typical users to design more realistic motivating scenario ?

Should we also explicit the set of personas like Polifonia ? For our case it would not be "a set of personas" but "a set of agents" with specifical features like :

• Human Agent / Artificial Agent
• Cognitive / Reactive
• Physical / Software
• ....

[andreiciortea]

@DrLeturc if we consider roles for agents where the roles are domain- or application-specific, I think this could relate very much to roles in multi-agent organizations and the regulation vocab. For instance, see the SemanticMOISE ontology developed by Madalina Zarafin, @Antoine-Zimmermann, and @oboissier [1]: https://www.emse.fr/~zimmermann/Ontologies/moise.owl

Not sure about defining a formal, domain-independent taxonomy for artificial agents in terms of reactive agents, cognitive agents, intelligent agents, etc.: it might be hard to achieve general agreement on these terms (e.g., what is really a cognitive agent?). I would go for it if we have motivating scenarios.

It could be useful to distinguish between human vs. non-human agents, we had this distinction in the STN Ontology [2] — and we'll find scenarios for the Social Web: https://w3id.org/stn/core

To summarize related consensus with @Antoine-Zimmermann and @oboissier for STN-Core:

• we distinguished between human vs. non-human agents (see stn:Person)
• we distinguished between physical and virtual artifacts; see stn:PhysicalArtifact and stn:DigitalArtifact
• we did not distinguish between physical or software/virtual artificial agents because we don't need this distinction if keep the separation of concerns between the agent's mind and the agent's body: an agent could have multiple bodies, which are artifacts and can be physical or digital
• we introduced the notion of stn:Role; in STN-Core, this was meant as a linking pin for a module that would describe normative and organizational concepts (what is now our "regulation vocabulary" in the HyperAgents project)
• a second linking pin to a future regulation vocabulary was stn:Group; roles and groups were the only concepts we found to be ubiquitous in models for organizations and institutions in MAS (i.e., the narrow waist / least common denominator we were able to identify)

---

References

[1] A.M. Zarafin, A. Zimmermann, and O. Boissier. Integrating Semantic Web Technologies and Multi-Agent Systems: A Semantic Description of Multi-Agent Organizations. Agreement Technologies 2012. http://ceur-ws.org/Vol-918/111110296.pdf

[2] A. Ciortea, A. Zimmermann, O. Boissier and A. M. Florea. Towards a Social and Ubiquitous Web: A Model for Socio-Technical Networks, WI-IAT 2015. http://andreiciortea.ro/publications/ciortea-et-al_WI-IAT2015.pdf

[DrLeturc]

A conclusion of this issue, after discussing with @FabienGandon.

We won't consider a folder "Persona" since :

• it seems very constraining for new members to go and look at each person created before being able to build their motivating scenario ;
• if we consider a folder "Agents" why not "Workspaces", "Artifacts", etc ? Maybe because "Personas" is related to users of the ontology and so only "Agents" would be relevant in such case..
• from the point of view of the methodology that we would like to put in place, it seems considering a folder "Agents" is too dependent of our project HyperAgent and it may be not relevant for other projects.