RE: Messaging Service Layer

A lot of these arguments are truly unresolvable since they are never in the context of "for what" -- meaning that we hear again and again that SOAP is doing things the wrong way, but we don't hear the "wrong way for what" part of the argument.

Here are some "for whats" it would be helpful to include in the arguments:

-- For XML in a LAN environment
-- For XML in a WAN environment
-- For XML in an EAI system
-- For XML is a B2B system
-- For XML as a replacement for HTML
-- For RPC-oriented systems
-- For message oriented systems

No doubt there are many ways to accomplish most tasks.  We have been hearing the same "SOAP isn't the right way" before the XML Protocols Workshop in Amsterdam in June 2000, at the workshop, and ever since.  

But it seems to be a classic case of each side making assumptions about their arguments that never get resolved, so we go on and on.

The use cases for REST and Web services are different.  The use cases for the Web and Web services are different.  The meaning of the term "application" is different in OSI than in TCP.

So can we please talk about "messaging for LANs" or "messaging for document exchange between message queueing systems" or etc. when debating the usefulness or appropriateness of given technology or architectural principles?  Otherwise this will continue, as it has till now, to be a very unproductive and unresolvable debate.  

What is anyone getting out of this other than the chance to re-state their positions yet again?

Eric

-----Original Message-----
From: Walden Mathews [mailto:waldenm@optonline.net]
Sent: Monday, February 17, 2003 10:58 PM
To: Champion, Mike; www-ws-arch@w3.org
Subject: Re: Messaging Service Layer



> > If there's concensus in the WG
> > about what you've written, I would request that this be
> > reflected in the arch doc so that IETFers and other OSI/IETF
> > model folks can be aware of the difference in terminology.
> > Or perhaps it could just go in the glossary, in the
> > definitions of "application layer", "transport layer", etc..
>
> There's nothing in the WSA charter or requirements about conforming to any
> flavor of the 7-layer networking model, AFAIK.  I'm unfamiliar with the
> IETF's position on this as a normative model for the highest-levels of the
> protocol stack, could you point me to that? I'd also be interested in
> hearing from anyone besides Mark who believes this is an important issue.

I think Mark is just asking the group to think hard before coding up
a spec that loses all the wisdom and helpfulness of these two imperfect
but very helpful models.  One way to "think hard" about it is to write
the text that Mark has requested above.

To the extent that credibility of the architecture document is important,
it's an important issue.

Walden

Received on Tuesday, 18 February 2003 09:31:16 UTC