[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [soa-rm-ra] Communicative Action Diagram
Frank and RA Team,
Couldn't agree more with Ken's statements. And we're
running out of time in my opinion. We cannot keep dragging this discussion
on endlessly. The PRD1 comment period has closed and we need time to
resolve issues. And, as Ken points out, he and I need to clean up and
converge the description and interaction models, and soon!
Collectively, we need to turn our attention to management and get closer to
our end state of wrapping this work up with a decent specification.
Personally, I do not like to notion of "Service Action" nor do
I like (as I've expressed before) the notion of "Communicative Action" or
other "types" of action, just as Ken noted. We agreed a long time ago
that for this RA, we were going to standardize on message exchange as the form
of Interaction. Can we just stick to Interaction? Sure there are
semantics associated with message exchange and that is what I was hoping you
would be focusing on in the Service Ecosystem view.
What needs to come through loud and clear is what in the heck
is Joint Action vs. Separate Action. And no more!
Keeping with the definition that Joint Action is the
application of intent by two or more participants in order to achieve a RWE, we
see that this translates to a service consumer (or consumer
agent) and a service provider (or provider agent, i.e., the
service) interacting in concert with each other with the goal
of achieving a RWE.* Separate Action, on the other hand, depends on
the participant's role. In other words, for a service consumer (or
consumer agent), the separate Action is sending a semantically and syntactically
correct message to a service provider (or provider agent, i.e.,
service) with the expectation that the service will respond by processing
that message and executing the necessary public and private separate
Actions (or operations) that results in a RWE
occurring.
*Yes, we need to address the dreaded "agents"
issue. If not in this view, certainly in the Realizing SOAs view.
That could be a point of discussion. Do we want to keep the Service
Ecosystem view at a very high level and only talk about service participants
(consumers and providers) in the abstract and not get into the details of
electronic communication (agents)? Open question. It's another issue
we have yet to resolve.
Hope we can come to some resolution on today's call my
friends! Would prefer that each member ask for the floor and be given time
without interruption instead of a "free for all," which I noticed after
listening to the MP3 from last week's call.
- Jeff
|
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]