W3C home > Mailing lists > Public > www-ws-arch@w3.org > October 2002

RE: Article on WS architecture and best practice ... may be of in terest

From: Vinoski, Stephen <steve.vinoski@iona.com>
Date: Thu, 3 Oct 2002 23:53:11 -0400
Message-ID: <4F4A31A61D72604FAF84C29C8EA28481567A64@amereast-ems1.IONAGLOBAL.COM>
To: "Prescod, Paul" <paul@prescod.net>, "Ugo Corda" <UCorda@SeeBeyond.com>
Cc: "Champion, Mike" <Mike.Champion@SoftwareAG-USA.com>, <www-ws-arch@w3.org>

Paul Prescod wrote:
> Ugo Corda wrote:
> >>this methodology
> >>defeats the whole purpose of Web services, which is to hide the
> >>implementation of a service completely behind an XML-based 
> interface. 
> >>VS.NET generates the interface from the implementation.
> > 
> > 
> > I don't see the conflict here. For any user of the Web 
> service the generated
> > interface does exactly that: hides the original implementation.
> > 
> > If the point made by the article is that Web services 
> interfaces should be
> > defined first and implementations should follow, this is 
> evidently not
> > possible in all those cases where Web services are used as 
> wrappers for
> > legacy implementations.
> 
> I think that the point is that the web service's interfaces should be 
> designed to make it into a good network application which could be 
> radically different than the appropriate interfaces for a 
> LAN-based or 
> desktop software component for all of the reasons described in the 
> "Waldo paper" and elsewhere. If you are just letting software 
> "generate" 
> your network interface from a pre-existing interface then the 
> chances it 
> is optimal as a network application is tiny.

Yes, I fully agree with you, Paul. I've described this phenomenon in several articles [1] [2].

Way back in the early days of CORBA (about ten years ago), a frequently-asked question was, "How can I turn my C++ object into a CORBA object?" The answer of course is, "You shouldn't do that." The granularity, type system, coupling, and design center for such an object are almost certainly all wrong for the CORBA domain. Ten years later, this FAQ still exists, but can now be applied to Java/.NET objects and Web Services too.

-steve

[1] <http://www.cuj.com/experts/1910/vinoski.htm>
[2] Vinoski, S. "Web Services Interaction Models, Part 1: Current Practice," IEEE Internet Computing, 6(3), May/June 2002, pp. 89-91. Available at <http://computer.org/internet/ic2002/w3toc.htm>.
Received on Thursday, 3 October 2002 23:54:49 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:09 GMT