RE: [wsrf] Singleton Resource Pattern

["Glen Daniels" <gdaniels@sonicsoftware.com>]

> It seems that whether it is an "implied resource" or an 
> "implied singleton" has nothing to do with 
> GetResourceProperty (GetProperty for that sake) operation. In 
> other words use of WS-Addressing does not add any semantic 
> value to the fact that one could retrieve a property by 
> sending a message to a service.

+1

This relates to the factoring question that came up at the initial F2F
in New Orleans.  In my opinion, the usage of WS-Addressing, or any EPR
mechanism, is and should be orthogonal to the WSRP patterns.  WSRP
should be, IMHO, about common patterns for state management for Web
services, and not get involved with the "where does the message go"
question.

There are certainly cases (WS-Notification, for instance) where it's
important to pass EPRs around, but even then all that is necessary is
that you be able to pass a serialized reference to a Web service.  That
might be as simple as a URL, and it might be as complex as a full-on
WSDL <service> (or WS-Addressing EPR) including feature/policy
requirements, dispatch headers, etc.  Obviously it takes more smarts for
an implementation to understand the latter kind of EPR, but I don't
think we need to mandate anything in that space.  We could do something
like BPEL has done, where we leave a placeholder for EPRs when they are
needed, and allow the community to settle into a single standard in a
little while after the W3C gets their addressing WG going.

I don't believe you should have to implement WS-Addressing, or any
particular dispatching mechanism, to get the useful commonalities
available in WSRP.

--Glen