- From: Savas Parastatidis <Savas.Parastatidis@newcastle.ac.uk>
- Date: Thu, 17 Jul 2003 16:51:14 +0100
- To: "Steve Graham" <sggraham@us.ibm.com>
- Cc: "David Snelling" <d.snelling@fle.fujitsu.com>, "Jim Webber" <jim.webber@arjuna.com>, <public-ws-desc-state@w3.org>, <public-ws-desc-state-request@w3.org>
> > This is an on going debate. > > I don't see SOA and Object Oriented at odds at all. When I worked up SOA > within IBM three or so years ago, there was absolutely no notion of > statelessness being a pre condition of a SOA. Furthermore, object > orientation was very much a part of the model, encapsulation, inheritance > etc. > Web Services follow the SOA model, no? The Web Services community has decided that services are stateless (WS-Arch document). So, that's the model we have to live with and build our solutions on. I happen to think that this model is the correct one (surprise surprise :-) You are probably referring to CORBA-based services. I agree that CORBA, in addition to the object-based architectural model that it defined at the beginning, it went on to deal with services-oriented architectures. However, it had to carry all that baggage of an object-based infrastructure for distributed computing and as a result failed miserably in cross-organisation-boundary deployments. This is where the Web Services model may be more successful. Why try to introduce CORBA-like features to Web Services before their usefulness is proved or disproved? Don't get me wrong. I like object-orientation a lot. However, I don't think object-based architectures are appropriate for a loosely coupled, inter-organisation environment. That's why I have expressed my concerns about OGSI and some of the new "features" that the Grid community is introducing into Web Services. .savas.
Received on Thursday, 17 July 2003 11:52:03 UTC