RE: SOAP breaks HTTP?

 Hi all, 8-)
 I have just written a message [3] [also quoted below] in which
I'm trying to analyze briefly what SOAP is and what it's
short-term course should be. Had I read any of this thread before
I posted that message, it would have been phrased differently,
but with the same basic idea.
 Best regards,

                   Jacek Kopecky

                   Senior Architect, Systinet (formerly Idoox)


Date: Wed, 27 Mar 2002 00:07:24 +0100 (CET)
From: Jacek Kopecky <>
Subject: What is SOAP?

 Hello all, 8-)

 it is now obvious now that serious disagreement comes from the
lack of the basic definition of what SOAP actually is -

 a) an extension of the web (or HTTP), or
 b) an RPC request/response protocol on top of *a transport 
layer*, or
 c) simply a messaging protocol with basically every message
being one-way (again, on top of a transport layer).

 I think the original Abstract Model [1] assumed (c), I think
that many of our audience still (unfortunately) assume (b), and I
think Mark Baker assumes (a) (at least I haven't noticed another
strong proponent of (a)). I believe that the majority expectation
is also (c).

 In W3C, it would be understandable if (a) was the goal. On the
other hand, the charter [2] for XML Protocol is more like (c).  
Let me quote:

"The goal is a layered system which will directly meet the needs 
of applications with simple interfaces (e.g. getStockQuote, 

 I'm not trying to say that (a) is the wrong course, on the 
contrary, it might be a big advance to the Web Architecture. 
Maybe the WS-Architecture and the TAG may come up with some 
proposals in this area.
 AFAIK, the history of SOAP is the transition from (b) to (c).  
I don't think there is a possible transition from (c) to (a)
other than starting from scratch. And I don't think (a) and (c)  
can be combined simply and nicely.

 Myself, I'm interested in where (a) could lead us, but I think
that for the time being we should agree on the approach here
described as (c), the messaging protocol, which may but need not 
use HTTP. The so-called tunneling approach, probably.

 (This is cross-posted to WS-Arch, so in your replies, please
consider if you want to keep the whole CC list.)

 Best regards,

                   Jacek Kopecky

                   Senior Architect, Systinet (formerly Idoox)


Received on Tuesday, 26 March 2002 18:26:57 UTC