Aftera rather circuitous intellectual journey, I find myself in violent agreementwith the bulk of what Savas lays out in his “ExplainingMEST” post. In many ways, I think he’s just giving a name to aset of ideas that a lot of the SOAP proponents have held implicitly for a longtime. Namely, the driving metaphor for the next generation of distributedcomputing architecture will be message exchange, not method calls. Myconfusion came from attempting to read too much in to what Savas was actuallysaying.

Like StefanTilkov, I originally thought that the MEST people were shooting for anapplication protocol. But as Savas points out, that’s not the goal ofMEST. It’s really about establishing a common model for distributedcomputing; just a REST lays out a model in which every resource is a containerfor state, MEST lays out a model where every behavioral entity works byprocessing messages. It’s a paradigmatic shift away from method calls –nothing more. The issue was that for at least some people, that paradigm shifthad already happened and confusion arose when it was given an unfamiliar name.

I remember Steve Swartz making a great point at one of the Indigo SDR’sabout this idea of driving metaphor. His point was that previous distributedcomputing architectures have tried to take a non-distributed metaphor – methodinvocation against a local object – and stretch it out to work acrossdistance. The new approach was quite the opposite – take thefundamentally distributed metaphor of exchanging messages with remoteservices and compress it down to work locally when needed. I think SteveSwand Savas are talking about basically the same thing.