Added as issue 913 (Clarification on workunit guard projection)

http://www.w3.org/Bugs/Public/show_bug.cgi?id=913

Begin forwarded message:

> Resent-From: public-ws-chor@w3.org
> From: "Gary Brown" <gary@enigmatec.net>
> Date: 7 October 2004 13:43:04 BST
> To: <public-ws-chor@w3.org>
> Subject: Clarification on workunit guard projection
>
>
> I understand the role of the globalizedTrigger, to provide a guard 
> condition for each role that is relevant to a work unit.
>
> However, if a work unit does not use a globalizedTrigger, and the 
> guard condition is related to just one of the roles involved in a 
> workunit, then does this mean that the condition at the other roles is 
> non-observable?
>
> In which case, does it have to determine what the first observable 
> action of the workunit would be, and monitor for it, to determine 
> dynamically if the workunit should be activated at that role? This 
> would mean that additional validation would be required to ensure that 
> the first observable activity (an interaction) was initiated from the 
> role that was associated with the guard condition.
>
> If this is true, then I think we need to add an additional rule 
> covering this situation, within the workunit section of the spec.
>
> Regards
> Gary
>