RE: choreography & orchestration must be defined in a context

Ugo
 
Semantics should really only be specified by the author of whatever is being
designed. So if someone is creating a WSDL definition for a Web Service,
say, then the semantics associated with that WSDL definition HAVE to be
understood before anyone can use it effectively. If they are not defined in
the WSDL itself, then either they need to be defined in a separate document
e.g.in HTML, or the user of the WSDL works out what to do using their
intuition or by trial and error and hope that they are right.
 
So what I really mean by saying that Semantics definitions are required, is
that it should be possible to include in the definition of what is being
designed (i.e. in a WSDL definition, a Choreography definition, etc.) some
text or other information that allows a user of the definition to understand
how to use the definition correctly with minimum chance of error.
 
I recognize though, that even if the Choreography Definition Language we
develop includes structures that allow the semantics to be included, there
are no guarantees that designers of Choreographies will actually define
them, or, if they do define them, they will be unambiguous.
 
Or to put it really simply, the Choreography Definition Language we develop
should be "self-documenting" by including XML elements and attributes that
allow explanations of the Choreography to be included in the definition
itself rather than separately.
 
David

-----Original Message-----
From: Ugo Corda [mailto:UCorda@SeeBeyond.com]
Sent: Sunday, November 30, 2003 7:03 PM
To: Burdett, David; Monica J. Martin
Cc: Jean-Jacques Dubray; Steve Ross-Talbot; public-ws-chor@w3.org
Subject: RE: choreography & orchestration must be defined in a context


> So without semantic definitions I don't see how interoperable solutions
could be built. 
> Secondly, a Choreography Definition Language that does not make it easy to
build 
> interoperable solutions is, IMO, useless and we may as well all go home!

Please keep in mind that all the standardization work done so far with Web
services (SOAP, WSDL, etc.) basically does not support any explicit
semantics to be associated with Web services definitions. Still it would be
hard to claim that it is not possible to build interoperable solutions with
them.
 
Ugo

-----Original Message-----
From: Burdett, David [mailto:david.burdett@commerceone.com]
Sent: Friday, November 28, 2003 5:26 PM
To: 'Monica J. Martin'
Cc: Jean-Jacques Dubray; Steve Ross-Talbot; public-ws-chor@w3.org; Ugo Corda
Subject: RE: choreography & orchestration must be defined in a context



Monica 

I am curious to understand why you say semantic definitions are outside the
scope of WS-Chor. Here's my thinking: 
1. Semantics are the meaning of things. 
2. Choreographies defined using the WS Chor language we create will contain
statements which will follow the syntax we devise.

3. Without including descriptions of the semantics of the statements in a
choreoraphy, the choreographies will not be understandable with any
certainty

4. But the choreography definitions MUST be understandable by everone who
builds solutions that will follow a choreography if each solution is to be
interoperable with every other.

So without semantic definitions I don't see how interoperable solutions
could be built. Secondly, a Choreography Definition Language that does not
make it easy to build interoperable solutions is, IMO, useless and we may as
well all go home!

Have I misunderstood? 

David 

-----Original Message----- 
From: Monica J. Martin [ mailto:Monica.Martin@Sun.COM
<mailto:Monica.Martin@Sun.COM> ] 
Sent: Wednesday, November 26, 2003 8:36 AM 
To: Ugo Corda 
Cc: Burdett, David; Jean-Jacques Dubray; Steve Ross-Talbot; 
public-ws-chor@w3.org 
Subject: Re: choreography & orchestration must be defined in a context 


Ugo Corda wrote: 

>Monica, 
> 
>Thank you for your Requirements quotation, but I am still not sure what to
conclude from that. 
>  
> 
mm1: 

    * Choreography 'of web services' only 
    * Bounding of the abstract level 
    * Semantic definition outside of WS-Chor 

These all seem to provide some boundaries of our scope, do they not? 

>Ugo 
> 
>  
> 
>>-----Original Message----- 
>>From: Monica J. Martin [ mailto:Monica.Martin@Sun.COM
<mailto:Monica.Martin@Sun.COM> ] 
>>Sent: Wednesday, November 26, 2003 8:00 AM 
>>To: Ugo Corda 
>>Cc: Burdett, David; Jean-Jacques Dubray; Steve Ross-Talbot; 
>>public-ws-chor@w3.org 
>>Subject: Re: choreography & orchestration must be defined in a context 
>> 
>> 
>>Ugo Corda wrote: 
>> 
>>    
>> 
>>>David, 
>>> 
>>>I agree with you about the existence of all those levels. What I am 
>>>not sure about is whether level 0 is in scope for this group, given 
>>>the fact that the WG is supposed to deal only with Web services. 
>>> 
>>>For instance, let's take the example of a choreography designed to 
>>>describe the interactions of multiple BPEL nodes (which is the case 
>>>that in my view provides the most value right now in a Web services 
>>>context). In that case, all the BPEL endpoints present WSDL 
>>>      
>>> 
>>abstract 
>>    
>> 
>>>interfaces and that is all that is required to describe their 
>>>interactions from the choreography point of view. 
>>> 
>>>Of course, we could go to a higher level (your level 0) and 
>>>      
>>> 
>>abstract a 
>>    
>> 
>>>whole set of choreographies that not only can describe that 
>>>      
>>> 
>>particular 
>>    
>> 
>>>set of BPEL endpoints but can also capture the interactions 
>>>      
>>> 
>>among any 
>>    
>> 
>>>other groups of BPEL nodes that differ from the first group only by 
>>>syntactic definitions of message vocabularies (same semantics). 
>>> 
>>>I understand there is value in this, I just am not sure we 
>>>      
>>> 
>>should go 
>>    
>> 
>>>that extra mile. I would be interested in hearing what 
>>>      
>>> 
>>other people think. 
>>    
>> 
>>> 
>>>      
>>> 
>>mm1: Is Level 0 within our scope as defined in our mission 
>>statement and 
>>evolving requirements document? 
>> 
>>Reference: 
>>Introduction in Requirements Document (previous version from 
>>SRT): "The 
>>description of interactions among Web Services especially 
>>with regard to 
>>the exchange of messages, their composition, and the sequences"...... 
>>......A choreography description is a multi-party contract that 
>>describes the external observable behavior across multiple clients 
>>(which are generally Web Services but not exclusively so) in which 
>>external observable behavior is defined as the presence or absence of 
>>messages that are exchanged between a Web Service and it's 
>>clients..... 
>> 
>>The only other comment is that originally we had quite a 
>>discussion of 
>>'human' clients, and I believe this was deemed or questioned 
>>to be out 
>>of scope. Perhaps we should clarify this in the introduction of the 
>>requirements document, in order to put this item to bed. 
>> 
>>Thanks. 
>> 
>>    
>> 
>>>Ugo 
[Burdett, David] </SNIP>