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

RE: Magic

From: Christopher B Ferris <chrisfer@us.ibm.com>
Date: Sun, 18 May 2003 11:56:14 -0400
To: www-ws-arch@w3.org
Message-ID: <OF0133D6EB.AAF7FE59-ON85256D2A.0048548B-85256D2A.00578BFD@us.ibm.com>

+1 To Eric's response.

Mark, I think that it is indeed unfortunate that you continue to 
characterize those of 
us on this list as having "not done our homework". I personally find it 
insulting and have said so
previously [1].

IMO, the "uniform interface" that has succeeded on the Web is not the 
complete set
of HTTP methods (GET/POST/PUT/DELETE, and 
the-ones-that-you-make-up-as-you-go-along[5].)

It is pretty much limited to GET. The number of Web resources that allow 
PUT 
and DELETE methods through is so frighteningly small that I seriously 
doubt 
these methods even factor into the equation of "what made the Web 
successful". 

Could we draw a conclusion that adopting a single universal method for 
retrieving a representation
of a resource is (abstractly):

representation GET (URI identifier) throws Exception

Sure... might not be a bad idea at all. Is this signature enough? I doubt 
it. There's really more to the 
interface than that. There's certainly a need for an optional security 
context. In fact, there's a bunch of 
stuff that is carried in (optional) HTTP headers that is not captured in 
this signature. Maybe, just maybe, 
it would be a useful excersize to explore this a little further, maybe 
even develop a processing model.
Ring any bells?

Further, as I have argued previously [2], I am not convinced that HTTP 
POST has
any set semantic meaning other than "here, chew on this". Certainly, in 
practice
as opposed to theory, POST means whatever the authority has chosen it to 
mean in 
the context of the resource(s) to which it applies (and yes, quite often, 
it is the plural
and *not* the singular because the thing at the end of the URI is a 
program not a
resource, typically a CGI script/program interface) (and yes, I am quite 
aware
of the fact that Roy says that these "suck", and they may indeed "suck", 
but that misses 
the point that in practice they probably outnumber the uses of POST that 
fully adhere 
to the constraints of REST, and they do work and have been successful in 
their own
right).

Still further, *you* continue to dismiss and/or avoid[3] any discussion on 
points 
that are well intended explorations[4] of the very claims you and others 
are making
when you assert as a foregone conclusion the hypothesis (and it can be 
nothing but an 
hypothesis because it has *not* been proven by any stretch of the 
imagination)  that
the same architectural constraints that apply to human-centric Web 
browsing 
when applied to program to program communications yield the same (desired) 

properties.

IMO, the members of the WSA and many others who participate on this list
*have been* making more than a *small* effort. I can certainly make that 
claim for
myself.

Mark, do yourself a favor, don't assume that it is *we* who "don't get it" 
and
don't assume that because we don't agree with you that it is because we 
"don't get it".

Thanks,

Christopher Ferris
Architect, Emerging e-business Industry Architecture
email: chrisfer@us.ibm.com
phone: +1 508 234 3624

[1] http://lists.w3.org/Archives/Public/www-ws-arch/2002Dec/0184.html
[2] http://lists.w3.org/Archives/Public/www-ws-arch/2003Jan/0351.html
[3] http://lists.w3.org/Archives/Public/www-ws-arch/2002Dec/0265.html
[4] http://lists.w3.org/Archives/Public/www-ws-arch/2002Dec/0263.html
[5] http://lists.w3.org/Archives/Public/www-ws-arch/2003Jan/0400.html

www-ws-arch-request@w3.org wrote on 05/18/2003 08:27:22 AM:

> 
> Mark,
> 
> This is a very interesting response.  You do not allow for the 
possibility that I (and presumably 
> others) might understand but still disagree. 
> 
> The lack of recognition for opposite viewpoints is usually a 
disqualifier for participation in a 
> public forum such as this, and I strongly urge you to either admit the 
possibility or withdraw. 
> 
> For the record, I completely agree that the Web is an historical 
success, and that REST as an 
> architectural description is very well suited to the Web.  I think 
uniform interfaces work well 
> for the Web, and I do not think the Web is either trivial or crude. 
> 
> But I also do not think that REST is an appropriate architecture for Web 
services.  The problem 
> space of program to program communications is sufficiently different 
from the problem space of 
> hypertext publishing to require a different approach.  I do not agree 
that tunneling is evil, and 
> I do not agree that WSDL interfaces are limiting, because I do not agree 
that WSDL is used in the 
> same way as REST, or should be.
> 
> The purpose of my email was to highlight the significance of factors 
outside of technical and 
> architectural purity.  Part of the argument I often hear about REST is 
that it has succeeded, 
> therefore it's good.  In the case of SOAP, WSDL, and UDDI the same is 
true, and that viewpoint 
> needs to be acknowledged.  Web services products do not implement REST, 
they implement SOAP, WSDL,
> UDDI, and a number of other specifications that are starting to emerge.
> 
> We, as a working group, have the choice of accepting the fact and 
working toward the goal of 
> defining an architecture that embraces them and the concepts they 
embody, or risking irrelevance.
> 
> Eric
> 
> 
> -----Original Message-----
> From: Baker, Mark 
> Sent: Saturday, May 17, 2003 2:49 PM
> To: www-ws-arch@w3.org
> Subject: Magic
> 
> 
> 
> On Sat, May 17, 2003 at 01:50:06PM -0400, Newcomer, Eric wrote:
> > Let's please forget about REST, the Semantic Web, and the other 
academic exercises and focus on 
> solving problems for business. 
> 
> Just because you don't understand it, please don't denigrate it.
> 
> The Web generates more business value every hour of every day, than
> SOA-style systems (combined) ever have or ever will.  Unconstrained
> interfaces are simply not suitable for the Internet.  Period.  End of
> story.  The empiricial evidence backs me up on this too.  The fact that
> lots of otherwise intelligent people (whose only mistake is they just
> haven't done their homework on Internet scale systems) don't understand
> that, does not all of a sudden make it so.
> 
> I *understand* that the Web looks like some trivial and cutesy
> human-driven system that enables crude UIs to be deployed in a thin-
> client fashion.  6 years ago, I thought the same thing too, and was a
> big fan of SOAs.  But it is *SO* much more, and I *WISH* people would
> just try to make a *SMALL* effort to understand that without getting all
> defensive and nervous about the prospect that they're mistaken; heck,
> if you are wrong, you're in good company 8-)  When I started studying
> the Web in 96/97, I wasn't expecting to learn what the Web was; I was
> just trying to find out what made it so successful so that I could
> incorporate that into my work (with CORBA, at the time).  I was as
> surprised as anybody to learn what I did; that not only did it have some
> neat things to offer CORBA, it removed the need for CORBA altogether.  I
> remember being absolutely stunned for about a week in May of '98 after a
> talk with Roy in which I finally "got it"; I hardly slept.
> 
> Anyhow, excuse me for venting.  I just couldn't let that tripe pass by
> without responding.
> 
> *PLEASE*, everyone, try to make an effort to understand it before you
> dismiss it.  If you understand it, but still want to dismiss it, then
> go nuts. 8-)
> 
> MB
> 
> 
Received on Sunday, 18 May 2003 11:56:27 GMT

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