[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [soa-rm-ra] Service participants
I've added my comments below. Let us know when the tags get to be too annoying to deal with. <original> The function of a system based on Service Oriented Architecture is to facilitate the interaction between participants; where this interaction is characterized as being in terms of providing services and consuming services. </original> <comment> A running thought I had is that this is not just for the benefit of the consumer or else the provider would have no motivation to be a participant. </comment> <suggestion> ... where this interaction is characterized as being in terms of providing services and consuming services to realize mutually desirable real world effects. </suggestion> <danny comment> agree </danny comment> <original> A participant is a human, corporation or non-human agent ... </original> <comment> How about collections of human and non-human entities, e.g. a car being driven on a highway or a claimant dealing with an insurance company? Not sure of the suggested words that follow, but something along these lines. </comment> <suggestion> A participant is an individual (either human or non-human functioning entity) or a group of individuals (where the group in certain contexts may be considered an aggregate individual) ... </suggestion> <danny comment> agree </danny comment> <original> The reason for this is twofold: an extremely common mode of interaction is where a provider participant offers some functionality as a service and a consumer participant uses that service to achieve one of his or her goals. </original> <comment> It says "twofold" but I don't see a second. </comment> <danny comment> twofold refers to producer and consumer </danny comment> <original> In normal parlance, the service provider commonly refers to either the ultimate owner of the capability that is offered or at least an agent acting as proxy for the owner. </original> <suggestion> Consider adding the following: For example, an individual may own a simulation program that encompasses a capability but will enter into an agreement with another individual (the proxy) to provide SOA access to that capability. </suggestion> <danny comment> I don't see the suggestion following the description for service provider. "Simulation program" lost me. </danny comment> <original> A service consumer is a participant that actively engages with a service... </original> <suggestion> For consistency with RM: A service consumer is a participant that interacts with a service... </suggestion> <danny comment> "actively engages" provides more meaning than "interacts". When would the start of a consumer/producer interaction not be due to the active engagement on the part of the consumer? </danny comment> <original> However, the service consumer normally refers to the ultimate beneficiary of the use of a service's capabilities. </original> <comment> This gets back to the comment above about mutual benefit. The suggestion changes the message but I believe is still accurate. </comment> <suggestion> However, the service consumer normally refers to the initiator of the use of a service's capabilities. </suggestion> <danny comment> If "actively engaged" stayed the same then "ultimate beneficiary" provides additional meaning. </danny comment> <original> Apart from the mechanical aspects governing how participants interact with and offer services, it is important to capture the relationships between participants that are mediated by services. At its most basic, this refers to the use of services to achieve a real world effect: a service is used by a service consumer in order to achieve a change in the real world that meets a desired goal. </orginal> <comment> Mutual benefit again. </comment> <suggestion> a service is provided by a provider and used by a service consumer in order to achieve a change in the real world that meets a desired goal. </suggestion> <danny comment> Either description reads ok to me. </danny comment> <original> A role is an identified relationship between a stakeholder and a social structure that defines the ... </original> <comment> Attributes effectively (and sometimes dynamically) define a "role" even if the role doesn't have a name and has never been previously defined. I don't necessarily want to get into a RBAC vs. ABAC discussion, but I think we need to capture this thought and not have this be read as assuming only predefined or static roles. </comment> <danny comment> agree </danny comment> <original> However, corporations, governments and even society, are abstractions: a government is not a person that can perform actions -- only people can actually do things. When a corporation performs an action it is always the case that some person or persons execute some action acting on behalf of the corporation. </original> <comment> I agree in principle but what about nonhuman agents? If I set up faulty logic and something results that I hadn't intended and possibly didn't want to occur, did I actually "do" it or just create the conditions under which it happened? (I don't even want to get into the future of machine intelligence.) I realize this is often a matter of law and liability, but I don't think we can say -- only people can actually do things. </comment> <danny comment> Agree. The machines one hundred years from now will love us for considering them. :) </danny comment> <original> This means that, when the persons acts appropriately, it is as though the other person or entity performs the action. In the case of non-human entities, the only way in which they can perform actions is via the empowerment of people. </original> <suggestion> Would it be helpful between the two sentences to add: An example of this is granting someone a power of attorney. </suggestion> <danny comment> Agree with example. <suggestion> Expand "empowerment of people" to something like "empowerment of an authorative entity." This goes back to Ken's last comment. </suggestion> </danny comment> __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]