ebxml-msg — archive

[System]

David, The diagram I drew IS the agreed upon behaviour of an IM w/r/t routing (e.g. IM routing is NOT an MSH function). It is almosgt straight from the white board in Tokyo (I think that was the meeting where we agreed on this). The diagram isn't in the spec, but it probably should be to help clarify what we meant. Note that I'm not hopping over any unreliable nodes in what I suggest at all. There are only reliable MSH nodes involved. Cheers, Chris David Fischer wrote: > > Chris, good explanation about how things might work. However, none of this is > in the spec. If this is all correct, then why bother to have > reliableMessagingMethod in Via -- or why bother having it at all? I don't > dispute that your way would work, and maybe we should do it this way, but what > you are describing does not align with the current spec although it could be a > subset of the current functionality. > > The way RM was designed was to allow for other RM methods. Your explanation > does not cover those and I can't find your "routing application" anywhere in the > spec. If this is some sort of source routing protocol (workflow routing), we > have specifically avoided that in the current spec. I have tried to stay out of > the IM functionality discussion but I have to think that an IM which does more > than store-and-forward is out of our scope for this version. > > I do like the way you hop over unreliable nodes. Let's just hop over everything > and do end-to-end RM only! > > Regards, > > David Fischer > Drummond Group. > >