W3C home > Mailing lists > Public > www-ws-desc@w3.org > June 2002

[amtf] my view of what the Abstract model should do and why

From: Jacek Kopecky <jacek@systinet.com>
Date: Wed, 5 Jun 2002 15:01:51 +0200 (CEST)
To: fablet@crf.canon.fr, <keithba@microsoft.com>, <ksankar@cisco.com>, <moreau@crf.canon.fr>, <prasad.yendluri@webMethods.com>, <sandkuma@cisco.com>, <sanjiva@watson.ibm.com>, <Waqar.sadiq@eds.com>
cc: Web Services Description mailing list <www-ws-desc@w3.org>
Message-ID: <Pine.LNX.4.44.0206051318590.20907-100000@mail.idoox.com>

 Hi all. 8-)
 When I was suggesting that we create an abstract model and when 
I was creating my first draft[1] of it, this was my motivation:

 The AM would be a short document containing all the
architectural terms in a clear and concise manner - for
clarification of the terms and their relationships, not for the
"big picture" encompassing the world outside WSDL. Using this 
concise document it would be easy to articulate and argue various 
design points and it would be clear what stuff we can do "as we 
like" because it doesn't matter from the point of view of the 
actual function of WSDL (like import).

 Thinking about the AM and writing it would highlight the
differences between WSDL 1.1 and what we think it should be and
let us think why there are differences.

 The AM would not solve issues like "what is the semantics of
import", it would, on the other hand, help solve (or at least 
clarify) issues like what MEPs are and which ones we should/must
support.

 If we decide we will continue finding issues with the spec and 
"patching" them (possibly involving some relatively minor 
rewrites), we will be dealing with a) editorial and syntax issues 
and b) design issues.

 The first group of issues we have to deal in any case but these 
issues are quite minor.

 The second group of issues are basically the differences between
the implied underlying model and what people think the model
should be. These issues will be harder to formulate (and even
find) without an explicit formulation of the underlying model 
because everyone has a kind of model inside their minds and these 
models may vary.

 I believe that without an explicit AM we will too often hear
Sanjiva (or other authors of WSDL 1.1) say "what we originally
meant was this: ..." in the middle of discussions where various
people clearly disagree about the intent of the spec. I'm not
saying it is wrong for Sanjiva to say that, I'm just saying that
it can be done beforehand in the AM and that the discussions can
be simpler later.

 And if we decide the AM is a good thing (in the form I 
suggested), we can proceed by taking the AM and creating a syntax 
for it from the scratch, having dropped the original spec. Of 
course we could (and we should) reuse most of WSDL 1.1 syntax in 
the process.

 Hope I've clarified it a bit,

                   Jacek Kopecky

                   Senior Architect, Systinet Corporation
                   http://www.systinet.com/

[1] http://lists.w3.org/Archives/Public/www-ws-desc/2002May/0153.html
Received on Wednesday, 5 June 2002 09:02:01 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:58:20 GMT