OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

soa-rm-ra message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: Re: [soa-rm-ra] Agent in Service Ecosystem View (Yes/No)?

I am sensitive to the people writing up security and what they may need said about agents, so anything I say below is secondary to those needs.  

In general, I'm not concerned with Agent as a proxy because it is unusual that we act strictly from our own motivation, i.e. I am usually a proxy in some sense for the policies and priorities set by my management.  For RA purposes, it is irrelevant who sets our goals but only that those are our goals during a particular interaction.  As a proxy, I may faithfully reflect management's goals or I may do so in a half-hearted fashion, but the system deals with what I do and not with what others wish me to do.

For the non-human agent, how about we have something somewhere to discuss the concept of the non-human agent and note that such an agent is always implied when a human is part of the interaction.  In circumstances where we need to say something about the agent, we make the agent explicit but consistent with the something somewhere discussion.

I'm not sure whether the something somewhere is in the Ecosystem View or the Realizing View.


On May 28, 2008, at 2:01 PM, Jeffrey A. Estefan wrote:

Looks like some progress was made on Action.  One issue that remains open is the notion of Agent and its use in the Business via Service View (or Service Ecosystem View or Using SOAs View or whatever we're going to agree to call it).
It seems that the way Agent is defined in the RA PRD1 (lines 570-571) implies that it is a non-human entity (although the definition does not state this) and it is an entity that is required for use in electronic communications.  Personally, I don't have a problem with this, provided we clarify that the agent is a non-human entity.  Another option would be to drop this type of agent from the Service Ecosystem View altogether and concentrate on the core Participants and not worry about electronic communications.  In this case, we would delegate the discussion of Agents (or Components) to the Realizing SOAs View where electronic communications comes to play.
That said, another interpretation of Agent could be (as the stereotype implies) a proxy for a Participant.  In this case, perhaps a Participant asks another entity to act on its behalf, independent of electronic communication.  This could be a human entity (or a non-human entity?).  This perspective is not really consistent with the Stakeholders and Participants model currently described in Section 3.1 in which Agent is an (implied non-human) entity that ties human Participants to the electronic communications world.
So we should decide whether to keep Agent in the Ecosystem view and firm up its definition, i.e., is it a non-human entity that ties Participants to electronic communication (as more or less currently defined) or Agent as a proxy that could be another entity (human or non-human) that acts on behalf of the Participant.  And once we decide, we need to clean up the supporting visual models and definitions of terms used throughout the Service Ecosystem View for consistency.  For example, the definition of Intent (lines 784-787) is very specific to agent and really should read "participant (or agent)" on all occurrences of agent.  Assuming we keep with the current definition of non-human agent.
 - Jeff


Ken Laskey

MITRE Corporation, M/S H305     phone:  703-983-7934

7515 Colshire Drive                        fax:        703-983-1379

McLean VA 22102-7508

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]