References to the idea(s) of robust one-way mep...

["Dale Moberg" <dmoberg@cyclonecommerce.com>]

David Orchard had a review of meps and the  wsdl, soap, ws-address situation that contains a diagram of robust in-only that is at least in the  ballpark of what we should support IMO

 

 

http://www.pacificspirit.com/Authoring/async/async-scenarios.html#oneway-robust-inonly

 

 

The only “fuzz” on this diagram pertains to the status of the “fault”. IMO, it should not be an application error (which should be reported in what JD was calling a but instead some middleware detectable error condition and this would include ebXML errors.) As I have suggested in other messages, I think we should try to be clearer in the text about the error distinctions and their placement in the “packaging” (the network protocol data unit’s format layers).

 

The WSDL 2.0 draft has a discussion of the concept of a robust in-only MEP

 

http://www.w3.org/TR/2005/WD-wsdl20-adjuncts-20050803/#robust-in-only

 

It is somewhat cryptic though.

 

 

That’s a start, anyway.