[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: [ebbp] 10/10/2004: BPMN Use in ebBP
On 7 Oct 2004 Thursday, we had a good session with the BPMN team re: support for business collaboration. Here are a few highlights of the discussion thus far: * Most BPMN Collaboration typically (most often) doesn't involve an intermediary managing the collaboration. It is peer-to-peer. * In the notation, evaluate need to recognize the BSI that is the interface for that part for the business collaboration. BSI verifies collaboration conformance to the description. o (paraphrase of St. Amand comment) Looking at the BSI, the endpoint is an abstract event, a boundary. The start, intermediate and end are instantiations of that boundary in BPMN. We actually need the abstract event. o [my comment] This to me represents a type of sorts. * BPMN can visualize: o The message flow and associate with a sequence (see comment below) o An event-based gateway (could support Decision for example) o Can represent: + Transactions (subprocess focused) [see comment below] + Events + Timers Discussion items: * Pending from 7 October o May need extensions for BT patterns. o Expressing bi-directional (capable of either side initiating) behavior [1] o Represent possibility of multiple requests and responses. o More on the abstract interface - Is it explicitly modeled? re: BSI. * Others [2] o Exceptions - BPMN allows an exception event. However, it is understood in the context of compensation. (May be more future item). o BPMN recognizes tokens that exist 'with' the message flow and appear to represent guards on a state transition. Do we need to consider or include? o What benefit could we get from using the 'sequence" element enforceable in BPMN? Could it be like a constraint? + BPMN considers a process without a pre-defined sequence to be ad hoc. Our patterns are anything but ad hoc but I am uncertain if the constraint is on execution in sequence or a design constraint that recognizes a sequence is at least expected. o Is a subprocess rich enough for composition use? We will meet again with BPMN in October. Team members are also welcomed to attend. Details will be forwarded when available. [1] See Change PO in example from White. This will be revised given the discussion. [2] My inputs References: 1. Our initial draft submission (example only): http://www.oasis-open.org/archives/ebxml-bp/200409/msg00008.html 2. White example 2 (BPMN): http://www.oasis-open.org/archives/ebxml-bp/200409/msg00078.html 3. Highlights of the BPMN spec relevant to ebBP (Thanks to Sally St. Amand): http://www.oasis-open.org/archives/ebxml-bp/200410/msg00030.html 4. BPMN session notes, 7 Oct (also available in Kavi Documents section): http://www.oasis-open.org/archives/ebxml-bp/200410/msg00045.html 5. Dubray updated diagram for BPMN (working draft): http://www.oasis-open.org/archives/ebxml-bp/200410/msg00042.html and http://www.oasis-open.org/archives/ebxml-bp/200410/msg00037.html
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]