Question about nomenclature (interaction model)

["Jeffrey A Estefan" <Jeffrey.A.Estefan@jpl.nasa.gov>]

Frank,

I've been studying how to describe any number of service interaction modes
but we need to converge on specific participants before I can embark on the
modeling process.  If we simply use "service consumer" and "service
provider" as abstractions for all consumer and provider entities (humans,
organizations, software agents, hardware agents, etc.) then describing the
interaction modes for various message exchange patterns (MEPs) is a piece of
cake.  This is the approach that seems to be commonly used in the industry
publications.  The obvious missing link here is "service" and often you'll
see service and service provider used in the same context, depending on the
perspective (business or architectural).  Service consumer and service
provider are obviously highly overloaded terms!

It seems for this RA we want to do more than continue to overload the notion
service consumer [/service requester] and service provider and that we'll
want to differentiate human and organizational roles from the pieces of
software and/or hardware that mechanize the functionality offered by
services.  We could choose to leverage the distinction you and your
colleagues made in the WSA or not (e.g., Requester Entity, Provider Entity,
Requester Human, Provider Human, Requester Agent, Provider Agent) with the
simple interaction shown on Fig 1-1.  Again, "service" is notably missing
but is defined in the WSA as an abstract notion whereas the agents are
concrete (allowing the swapping in an out of agents but providing the same
service functionality).  Obviously, these notions were created to help
resolve some of the obvious ambiguity surrounding the overloaded terms of
requestor [consumer] and provider.

Perhaps we will want to use a different nomenclature to help resolve some of
these ambiguities.  Hopefully, we will be able to come to converge on the
terminology soon and then I'll develop the interaction models.  In the
meantime, I'll focus on the information and behavior model elements of this
architectural view (Interaction View) focusing primarily on the former with
a fair treatment of metadata.  If we choose to adopt to describe this RA in
terms of architectural views (and I hope we do!), then I'd suggest that the
Visibility and Real World Effect sections become part of the Interaction
View.

Comments from the rest of the RA team on this thread are welcome.  For
reference purposes, a copy of the W3C WSA is available on our Kavi site at
the following URL:

http://www.oasis-open.org/apps/org/workgroup/soa-rm-ra/download.php/17572/wsa.pdf

Regards...

 - Jeff