RE: [wsbpel] Issue 37 - Proposal for vote

["Furniss, Peter" <Peter.Furniss@choreology.com>]

Are you thinking of a "singleton" process, where there is only one
instance of the specified process ? Such a process wouldn't use
correlation sets as an "alias for the identity of the business process
instance" (10.1, penultimate paragraph), though it could use them for
linking its replies to the inbound notifications. But I'm not sure
singleton processes are envisaged.

If it isn't a singleton, then the inbound message is "trying" to get to
some specific, already running, instance (since it is
createInstance="no", so there needs to be something in the message,
including any headers and protocol control information from lower
layers, that allows the engine to work out who it is for. That
information appears in the process script as the correlation set(s),
initiate=no.

If singletons are allowed, then I guess the rule needs modifying - or
you could warp it the other way, and say that a singleton has an
invisible correlation set on all its activities ( with a value
<reality>one</reality> ?)

But as Satish points out, there may be other problems then.

Peter

> -----Original Message-----
> From: Assaf Arkin [mailto:arkin@intalio.com] 
> Sent: 17 June 2004 21:56
> To: Furniss, Peter
> Cc: Satish Thatte; Alex Yiu; wsbpel@lists.oasis-open.org; 
> Yuzo Fujishima; Dieter Koenig1; Yaron Y. Goland; Nickolas Kavantzas
> Subject: Re: [wsbpel] Issue 37 - Proposal for vote
> 
> 
> Furniss, Peter wrote:
> 
> >  
> > RULE: For an inbound activity with createInstance=no, there 
> must be at
> > least one correlation set with initiate=no to allow the 
> message to be 
> > routed to the appropriate process instance. There may be other 
> > correlation sets with yes or rendezvous:
> >  
> 
> How would that work if the process wants to receive 
> notifications that 
> are broadcast to a large audience (pub/sub)?
> 
> Assaf
>