W3C home > Mailing lists > Public > www-ws-arch@w3.org > May 2003

Re: Separate concepts for "service" and "targetResource?" (was RE: /service/@targetResource ?)

From: Walden Mathews <waldenm@optonline.net>
Date: Thu, 22 May 2003 09:53:48 -0400
To: David Orchard <dorchard@bea.com>, "'Assaf Arkin'" <arkin@intalio.com>
Cc: "'Ugo Corda'" <UCorda@SeeBeyond.com>, "'Champion, Mike'" <Mike.Champion@softwareag-usa.com>, www-ws-arch@w3.org
Message-id: <002001c32069$930a8160$1702a8c0@WorkGroup>

> > There's a reason why procedural programming was kinda dropped
> > a few decades ago in favor of the object model.  I have a feeling we
> > are going to see that metamorphosis again soon.  This is like taking
> > one object method, promoting it to global scope, and then trying to
> > bind it back to only a single instance.
> >
> >
> Indeed.  Because Object-orientation is fundamentally broken when it comes
> real world things.  The "subject", ie the object performing the "verb",
> the "object", ie the real world thing, are separate.  In order to talk
> real world things, subjects and objects - objects in the language sense -
> need to be first class citizens of the grammar.  You need to have the
> subject,verb,object structure to be able to make proper assertions about
> things are and who can do what to the things.  WSDL is making some simple
> assertions, what services an object exposes.

WSDL seems like the wrong place to try to solve that problem.

> Now you can argue that O-o can model this by having an o-o object
> the real-world object.  But then there's all these darned leaky
> problems that we spent 25% of our time in software dealing with.  And in
> fact, WSDL also supports this model because you don't have to provide the
> resource URI.  WSDL has the classic problem that the assertions about the
> resource may be in different structures and owned by different people, and
> think they solve it in a nice way.

I agree that OO metaphor sometimes gets ridiculously out of hand.
My favorite is the Sandwich object that eats itself.  But I haven't
known that problem to hamper implementation.  The bits can still
end up in the right state, even if we're flaky in the way we describe

See, I think the OO model works fine for exposing what needs to be
exposed and hiding what needs to be hidden, in order to enable service
of all kinds.

And I suppose, worst case, the services for which target URI are
specified are just static singleton objects out there.  If they're
then we obviously don't care about scaling them up, but they're kind
of marginal in the scope of internet applications at the same time.

Received on Thursday, 22 May 2003 09:49:34 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 23:05:52 UTC