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

Re: Normative constraints on the WSA

From: Walden Mathews <waldenm@optonline.net>
Date: Sun, 18 May 2003 15:53:29 -0400
To: Christopher B Ferris <chrisfer@us.ibm.com>, www-ws-arch@w3.org
Message-id: <008001c31d77$27bfc9a0$1702a8c0@WorkGroup>

Christopher,

My mistake.  See [1].  You said RPC was implied.

Notwithstanding that, I'd been hearing mostly anti-RPC sentiment on
this list, and today I'm seeing a resurgence of RPC interest, so I'm
wondering what that means.

Walden

[1] http://lists.w3.org/Archives/Public/www-ws-arch/2003Apr/0188.html
----- Original Message -----
From: "Christopher B Ferris" <chrisfer@us.ibm.com>
To: <www-ws-arch@w3.org>
Sent: Sunday, May 18, 2003 1:26 PM
Subject: Re: Normative constraints on the WSA


>
> Walden,
>
> You said:
>
> > I'm not saying you're wrong here; you've been around it way longer
> > than me, but I had the strong impression that EVERYONE at this
> > point felt that RPC was dead as the principal pattern of Web Services.
> > Recently I asked Chris Ferris whether some WSA language ought to
> > include reference to RPC, and the anwer was 'no', for example.
>
> I'd be very interested to know when/where I said that WSA should not
> include
> reference to RPC. Maybe in a specific context within the document, but not
> as a
> generalization.
>
> Certainly, I would agree that a message- rather than procedure-oriented
> approach is superior ;-), but I don't believe that I ever suggested that
> WSA not
> accomodate its application. I think that despite the short-comings of
> RPC-oriented
> systems, they'll be around for a long time to come.
>
> Cheers,
>
> Christopher Ferris
> Architect, Emerging e-business Industry Architecture
> email: chrisfer@us.ibm.com
> phone: +1 508 234 3624
>
> www-ws-arch-request@w3.org wrote on 05/18/2003 12:57:55 PM:
>
> >
> >
> > > This current generation of Web services technology uses (abuses?) the
> Web.
> > > But it is NOT the Web. It is RPC-oriented middleware -- it is
> > > service-centric rather than resource-centric. It's about verbs rather
> than
> > > nouns. If I recall correctly, the folks that originally came together
> in
> > > April 2001 to talk about Web services and that recommended the
> immediate
> > > formation of this group weren't even thinking about REST at the time.
> We
> > > were thinking about RPC. And we wanted to define an over-arching
> > > architecture for this type of middleware.
> > >
> > > I think that's what this group should focus on.
> >
> > I'm not saying you're wrong here; you've been around it way longer
> > than me, but I had the strong impression that EVERYONE at this
> > point felt that RPC was dead as the principal pattern of Web Services.
> > Recently I asked Chris Ferris whether some WSA language ought to
> > include reference to RPC, and the anwer was 'no', for example.
> >
> > Why the current interest in "service orientation", by the way?  It
> > seems counter to the almost ubiquitous revolution from procedural
> > programming application models to object orientation.  What service
> > can a "service oriented" application provide that a "resource oriented"
> > one cannot?  I would say they are both about "action", and I'd be
> > interested to know whether people see one model subsuming the other,
> > and if so, which model that would be.
> >
> > >
> > > At the same time, I think that it would be an excellent endeavor to
> work
> > on
> > > the next generation of Web services -- a RESTful version of Web
> services.
> > > I'd love to see another Working Group started to focus on this work. I
> > just
> > > don't think that this work should interfere any further with the
> immediate
> > > work at hand.
> >
> > At the level where end clients would have interest, would these two
> > groups be solving different problems, or solving the same problem?
> >
> > > Most of the retail commerce success is based on CGI/ASP/JSP -- which
> very
> > > definitely tunnels method calls through HTTP. It isn't RESTful.
> >
> > ?? JSP is a convenience for Java HTTP Servlet, which is typically
> > set up to GET and POST for all operations, but can easily be configured
> > so the application can also use PUT and DELETE.  Where is the
> > tunnel?
> >
> > Walden
> >
>
>
Received on Sunday, 18 May 2003 15:49:11 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:19 GMT