RE: Proposed Choreograph Working Group Charter


There are two problems here:
1. Defining the semantics
2. Defining how to represent those semantics unambiguously and in a formal
rigorous way

For Business Documents, the first (defining the semantics) is going on in
some groups already e.g. UBL at -
there are other initiatives, but this one seems to be cross-industry.
UBL and others are defining their business documents using XML Schema
combined with rigorous conventions on how to document them (UBL is also
working on this) - XML Schema is too flexible to be used without

Business processes/choreogrpahies though are way behind. We do not even have
an agreed way of formally defining a choreography. I think that one of the
objectives of the Choreography WG will be to develop formal methods for
defining them.

Once we know how to define choreographies, there is still the problem of
defining the actual choreographies to use. 

Without both business documents AND choreographies, the chances of
interoperability are slight. For example, if everyone uses different
business documents (e.g. for orders) and different choreographies (i.e. the
sequence in which those documents are exchanged), then the widespread
interoperability requied to make eCommerce work will simply not occur.

On the other hand one size does not fit all.

There's still a LOT of work to do ...


-----Original Message-----
From: Francis McCabe []
Sent: Monday, September 09, 2002 1:48 PM
Subject: Re: Proposed Choreograph Working Group Charter

I cannot help thinking that if the group seriously considers being able 
to do business without semantics a positive rather than negative 
feature of today then there is some education necessary. If you are 
Ford, or Shell, it may be that you are able to bully your suppliers 
into a single accepted standard for the meanings of terms; however, 
even there there is a semantics -- it may not be written down, it may 
not be negotiable, and if you are not Ford etc. you may not have a say 
in it (and there get a chance to influence it) but hey you STILL have a 

The difference between a written down semantics and a 
back-of-the-envelope semantics is not that there is no semantic 
attachment to verbs such as "open_this_account"; its simply that the 
meaning is opaque and inherently fragile. (If you write both the 
customer and supplier `agent' then this may not matter very much; but 
if I write the customer agent and you write the supplier agent, then we 
begin to get into trouble fast.)


On Monday, September 9, 2002, at 01:36  PM, Cutler, Roger (RogerCutler) 

> It's not clear to me that "business semantics" is even in the scope of 
> the
> W3C.
> -----Original Message-----
> From: Champion, Mike []
> Sent: Monday, September 09, 2002 2:45 PM
> To: ''
> Subject: RE: Proposed Choreograph Working Group Charter
>> -----Original Message-----
>> From: Ugo Corda []
>> Sent: Monday, September 09, 2002 2:34 PM
>> To: ''
>> Subject: RE: Proposed Choreograph Working Group Charter
>> Yet another comment on the Charter:
>> I am wondering whether the Charter should say something about
>> addressing business semantics.
> My personal opinion [speaking as a member not a co-chair] is that this
> Charter will have a better reception if we constrain it to  a subset of
> "choreography" -- small enough to have existing practice to learn from 
> and
> so that a new spec can be written quickly, but large enough provide for
> technology implementing the spec to have a real benefit.
> So, I'd say "no" -- let's try to focus on technology that works across
> different business semantics (and even architectural styles).
>> while BPSS addresses business semantics most
>> of the other specs don't (that being considered a drawback of those
>> other specs).
> It may be a drawback for those other specs in terms of functionality, 
> but as
> a practical matter I doubt if very many people in the W3C or the member
> companies who might sign up for a Choreography WG believe that 
> "business
> semantics" is something ripe for standardization.

Received on Monday, 9 September 2002 17:12:28 UTC