- From: Cutler, Roger (RogerCutler) <RogerCutler@chevrontexaco.com>
- Date: Mon, 12 Jan 2004 20:13:20 -0600
- To: "Champion, Mike" <Mike.Champion@SoftwareAG-USA.com>, www-ws-arch@w3.org
I personally think that the reasons it is unwise to try to implement OO in SOA are more fundamental than that. I think that I have heard some rather well-known people saying the same thing very publicly. I don't know how to put it gracefully, but the basic idea, I think, is that there is too much detail you have to get right to implement OO remotely. You can do it, but it's gonna be fragile unless you have detailed control of the environments on both sides. Yeah, Microsoft can do it with .Net Remoting, but as I understand it their long term strategy is Web services, not .Net Remoting. The SOA approach brings robustness by agreeing to handle things at arms length and not to try to do too much. Who's that guy who used to work for DevelopMentor and now works for Microsoft? The guy that has something of a cult following? I think that's what he's been saying, more or less. I personally would be very happy to be prescriptive here. SOA good. OOP good. Best keep in separate tents. An architecture is not meant to describe everything that ever has existed or might exist. It is a recommended structure that includes some things and not others. On such a fundamental issue as this I see no reason to be apologetic about supporting this and deprecating that. And I don't think that in doing so that one is inherently deprecating grid computing. And again I remind you -- is this a lamb? Might as well go for the sheep, eh? -----Original Message----- From: www-ws-arch-request@w3.org [mailto:www-ws-arch-request@w3.org] On Behalf Of Champion, Mike Sent: Monday, January 12, 2004 7:46 PM To: www-ws-arch@w3.org Subject: RE: Proposed replacement text for Section 1.6 > -----Original Message----- > From: He, Hao [mailto:Hao.He@thomson.com.au] > Sent: Monday, January 12, 2004 7:32 PM > To: 'Champion, Mike'; www-ws-arch@w3.org > Subject: RE: Proposed replacement text for Section 1.6 > > I think we all agree "web services can be used to implement > OO as well as SOA architectures", but we also want to say > "web services are really intended for SOA. Web services for > OO are not recommended." How about "web services are well suited for SOA; Web services implementations of distributed OO across trust domains and when underlying transports are slow/unreliable are beyond the currently proven state of the art." I'd appreciate hearing from those who are investing heavily investing WS-Security, WS-ReliableMessaging, etc. to advance the state of the art here. Do you folks think this would make WS suitable for internet scale distributed objects, or that they are needed for secure, reliable SOA impelemtations?
Received on Monday, 12 January 2004 21:13:33 UTC