To resurrect a dead horse (as we seem wont to do from time to time) "communication models...cannot be architectural principles"
>From Fielding's dissertation: "This chapter introduces and elaborates the Representational State Transfer (REST) architectural style..." Interaction constraints are fair game. SOA specifies zero interaction constraints. WOA says that the interactions must be RESTful. IMO, this is just as reasonable within an architectural style as indicating that major components are "services" and must present "separately standing interfaces." Further, even if REST was an implementation choice (I do not believe that it is) an architecture/style placing constraints on implementation choices (e.g. the use of WWW technologies constraint of WOA) is not out of bounds, IMO. Just as not everything within an architecture will be SO, not all interactions will be RESTful. Thus, portions of an architecture will be SO, some will be WO and others will be something else. And that's okay, IMO, because style purity is unachievable and ultimately unnecessary. To reiterate from the other thread (about Chrome, of all things!): As creators of the term, WOA means what Gall/Gartner says it means! :-) -Rob --- In [email protected], Michael Poulin <[EMAIL PROTECTED]> wrote: > > I disagree that SOA "concept is finally understood and gaining > traction". Here are a lot of things to gain, especially, in the > business and enterprise areas. > > Setting WOA or Web Services as a "substyle of SOA" has two > troubling aspects: > 1) SOA is still taken as a technology (otherwise Nick Gall mixes > apples and oranges), we it is not > 2) communication models - Web Services/WS*-, REST, MOM/Event- > Driven - cannot be architectural principles, its are just > communication models; depending on the task one of them might be > more preferable than others. > > I agree with ZapThink in that SOA and WOA are abstractions of > different levels. > However, I disagree with using expressions such as Web-Oriented SOA, > Event-Driven SOA, Model-Driven SOA, Process-Driven SOA, etc. I > think it would be a mistake to limit SOA even in one enterprise to > only one communication mechanism ( unless it is very a simple > business where enterprise operates). I feel it would be more > appropriate saying something like "a SOA subsystem > with Web-Oriented communication model (WOCM)" > > > > - Michael
