- From: Mark Baker <distobj@acm.org>
- Date: Mon, 27 Oct 2003 09:16:17 -0500
- To: Savas Parastatidis <Savas.Parastatidis@newcastle.ac.uk>
- Cc: www-ws-desc@w3.org
Hi Savas, On Mon, Oct 27, 2003 at 02:53:44AM -0000, Savas Parastatidis wrote: > > You claim there's no operation, but there really is one, because there > > remains a notion of "success" and "failure". You just have to ask > > "success and failure of what?". The answer to that question is your > > operation, whether or not you choose to include it in your message. > > > > There doesn't necessarily have to be an answer. I can imagine situations > where there isn't a requirement for a response. I didn't say a response was required, only that the operation can fail or succeed. How that success or failure is signalled is another matter. > > So it's true that you can get a whole lot done with just the "process > > this state" implicit operation in your example - heck, EDI was built > > around that operation (AFAICT). But at least one more would be nice; > > an operation for *retrieving* the state of things. > > > > I don't use the terms "process" or "state". Sending a message is not an > instruction to an agent to do something. It's just a message. No other > semantics attached. Do you want to call that message "state"? That's up > to you. Ok, perhaps there's a nomenclature issue. We're simply talking about the difference between the following two things; <orderForm1> <carOrMotorcycle>car</carOrMotorcycle> <colour>blue</colour> <cc>1800</cc> <model>blabla</model> </orderForm1> and POST some-uri HTTP/1.1 Content-Type: application/order-form+xml <orderForm1> <carOrMotorcycle>car</carOrMotorcycle> <colour>blue</colour> <cc>1800</cc> <model>blabla</model> </orderForm1> You and I both want those to mean the same thing, i.e. "please process this document". I claim the latter approach has some advantages over the former in terms of extensibility and self-description. But we could build large scale systems with either approach, because integration complexity scales as O(N), rather than O(Nlog(N)) or O(N^2) with typical SOA with service- specific operations. > So, I am talking about message exchange within > this model and not the model described by REST. That *is* the REST model. We might use different terms and a different approach, but we're describing the same thing. Really. Mark. -- Mark Baker. Ottawa, Ontario, CANADA. http://www.markbaker.ca
Received on Monday, 27 October 2003 09:20:02 UTC