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

RE: Proposed Venn Diagrams

From: Cutler, Roger (RogerCutler) <RogerCutler@chevrontexaco.com>
Date: Fri, 9 May 2003 09:17:19 -0500
Message-ID: <7FCB5A9F010AAE419A79A54B44F3718E026EF6E9@bocnte2k3.boc.chevrontexaco.net>
To: "Champion, Mike" <Mike.Champion@SoftwareAG-USA.com>, www-ws-arch@w3.org

WOW!!  I don't know about anyone else, but I personally find these
diagrams INCREDIBLY useful.  Thank you.

One nit -- or rather question.  Why do you have no overlap whatsoever
between D and C?  I thought that if you are careful enough you can
define thingies in D that make people in C happy.  Isn't that partly
what the alligator wrestling is about?

-----Original Message-----
From: Champion, Mike [mailto:Mike.Champion@SoftwareAG-USA.com] 
Sent: Thursday, May 08, 2003 11:20 PM
To: www-ws-arch@w3.org
Subject: Proposed Venn Diagrams


I gave myself an action item to produce a Venn diagram to accompany
section 1.6 (the text we've been talking about for about 24 hours)  It
is attached in PNG format.

The first captures what I *think* is the consensus from the previous
thread and the call:

SOA is the all-encompassing concept for all this stuff.   I guess
"distributed systems" would be even broaded, and it would include things
like Unix RPC that don't attempt to define any kind of "service" model,
but let's not go there.

The Web IS-A instance of an SOA.  "Uniform Interface Services" aka REST
is a subset of the Web.

Since I don't want to argue about terminology right now, I gave the sets
labels, so "D"  (what people generally think of as "web services") a
subset of SOA and overlaps the Web, but is not a subset of the Web.
"SOAP over HTTP" is a subset of both the Web and Open Interface
Services, but there are other services that only partially use the Web
(e.g. an HTTP gateway into a mainframe service), and some that don't use
the Web at all (e.g COM and CORBA).

I think the scope of the WSA [warning: this goes beyond anything the WG
has discussed] is roughly set "F". (It might be called "XML Interface
Services", but let's not go there now).  That overlaps the REST circle
because one can (e.g. using SOAP 1.2's GET feature) build RESTful web
services; there was some dispute about that today, but the TAG settled
that a year ago and I ain't a gonna get into that swamp and
alligator-wrestle Roy Fielding, no sir-ee!!  It obviuosly overlaps the
"Open Interface Services" circle.  

The second diagram expands the "F" set to indicate that the scope of
this WG maximally includes services that use SOAP (F1), services defined
by WSDL (F2), services that employ both (F12) and services that use XML
but neither SOAP or WSDL (F0).  Perhaps it would be useful to do a poll
as to where people stand on the "in scope" question for these sets:  I
personally (not wearing chair hat or Software AG hat for that matter)
think that all of Set F is in scope.  Andrew Layman seemed to imply that
he thinks Set F1 should be the scope and I believe I've heard the IBM
folks imply that Set F2 is the scope ... and some hard-core people might
say that F12 is the scope.  I believe the Gartner Group defines the F1 U
F2 as "web services" (actually they put UDDI in there too).

Anyway,  

Thoughts?  Is this useful? Would the first of these diagrams be useful
in the WD we publish next week? I'm kindof sorry we didn't draw this
diagram and have this argument months ago, but maybe, just maybe we can
get a working consensus at the F2F. 
  
 
p.s., In case it's obvious, yes I *am* colorblind :-)  Suggestions for
improving these diagrams are welcome!  I'll attach the PPT in case
anyone wants to fool with them.
Received on Friday, 9 May 2003 10:17:47 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:18 GMT