[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: RE: [bpel4people] BP-92 Proposal
Hi Dieter, I see it as follows: We have two concepts..1) Routing Patterns 2) Composite Tasks We have two syntax versions which we can recommend to describe either one: 1) the complex people assignment to describe the complex "who" and 2) the composite/subtask syntax to describe the complex "what". There is certainly an overlap for what can be expressed with either syntax, but there is also a delta: - complex people assignment can't handle task definition of different types - composite/subtask syntax can't describe the dynamic determination of the number of to be instantiated sub tasks based on the evaluation of a logical people group and the number of resolved people (multiInstance) (assuming TC agrees to proposal 97 to remove the "multiInstance" from the <subtask> element) However, both, the complex people assignment and the composite/subtask syntax result in the same runtime instantiation model, that means for example that a Task Engine automatically instantiate all the subtasks, described by a complex people assignment as soon as the wrapper task moves in status "in progress" (assuming TC agrees on proposal 94) Regards -Oliver -----Original Message----- From: Dieter Koenig1 [mailto:dieterkoenig@de.ibm.com] Sent: Mittwoch, 22. Juli 2009 14:49 To: Kieselbach, Oliver Cc: bpel4people@lists.oasis-open.org Subject: Re: [bpel4people] BP-92 Proposal Hi Oliver, IMO, it would help to be more specific about when precisely routing patterns can be used as a special syntax for composite tasks. For example, something like: <spec-text> Routing patterns can be used as a convenient syntax to decribe the "who" in a composite task, where all of the following is true: - all subtasks have the same interface - all subtasks have an activation pattern of "automatic" - (whatever else is needed) </spec-text> You might then drop the later sentence "A composite task, including subtasks of different task types, can be desribed only using the composite task syntax". As a side remark, we need to clean up the usage of the term "task type" -- it is sometimes used in the sense of "task definition" (as opposed to "task instance") and sometimes in the sense of "task vs. notification" -- I will open a separate issue for this. Kind Regards Dieter König Senior Technical Staff Member, WebSphere Process Server Architect IBM Software Group, Application and Integration Middleware Software WSS Business Process Solutions Phone: +49-7031-16-3426 IBM Deutschland (Embedded image moved to file: pic01831.gif) E-Mail: dieterkoenig@de.ibm.com Schönaicher Str. 220 71032 Böblingen Germany IBM Deutschland Research & Development GmbH / Vorsitzender des Aufsichtsrats: Martin Jetter Geschäftsführung: Erich Baier Sitz der Gesellschaft: Böblingen / Registergericht: Amtsgericht Stuttgart, HRB 243294 From: "Kieselbach, Oliver" <oliver.kieselbach@sap.com> To: <bpel4people@lists.oasis-open.org> Date: 21.07.2009 18:57 Subject: [bpel4people] BP-92 Proposal Please check the comment (the last comment) in the Jira issue for the resolution proposal: http://www.osoa.org/jira/browse/BP-92 Thanks -Oliver
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]