Tomas Restrepo poses the thought that Web Service proxies hides what is happening under the covers too much, and so leads to confusion between the programming model and the messaging model for communication with a Web Service.
While this is certainly true in many respects, I actually think that is a good thing overall - hiding the "plumbing" and providing a higher-level abstraction (ie increasing developer productivity) must be better overall?
I think the main point is we (ie. the industry) need to do a much better job of explaining how the notions of client programming models, messaging models, and transport interaction style are almost totally orthogonal to each other.
For example, an RPC-style client proxy can work over the top of two completely separate message interactions under the surface (such as pairing of a WSDL one-way and a WSDL notification interaction), or even running as a WSDL request-response interaction over an asynchronous transport (by waiting for a correlated reply message on another messaging queue, for example), or use document/literal under the covers.
Once for start looking into it, there are some really cool possibilities here!
Cape Clear for one is doing some interesting things in this area, and I am sure others are too.
I think we will see many more "Smart Proxies" coming along very soon.
Copyright (c) 2008 Jorgen Thelin. All rights reserved.
This weblog is licensed under a
Creative Commons License.