Comments on: BPMN 2.0 from Visio Premium 2010

By: bruce

bruce — Mon, 29 Nov 2010 22:46:35 +0000

It is not a plugin or stencil. It is an xslt transform of a BPMN model created using Microsoft Visio Premium 2010′s BPMN stencil.

By: siyam

siyam — Thu, 04 Nov 2010 14:50:01 +0000

Hi Bruce,

Would you be willing to post your BMPN 2.0 Visio Stencil/Plugin?

By: What is a “Process”? « BPMS Watch

What is a “Process”? « BPMS Watch — Mon, 18 Oct 2010 23:25:06 +0000

[...] the same thing many long-time flowcharters think it means, as evidenced in a recent comment thread here. Blame for this lies mainly with the BPMN spec itself, which adds to its sins of omission by [...]

By: bruce

bruce — Sun, 17 Oct 2010 15:11:42 +0000

It’s not about the XML. The XML is just the storage format. It’s about what things like “process” and “pool” and “sequence flow” actually mean in BPMN. I think I need to post about this topic again. BPMN 2.0 had an opportunity to remove the confusion but the team stubbornly refused to do so. (I gave up after months of trying.) Anyway, I will take it up again on a new post.

By: Connie

Connie — Sun, 17 Oct 2010 14:03:55 +0000

Bruce,
Thanks so much for the explanation. I had thought the connection from Activity 1 to Activity 4 was satisfied by the message flow down and back from the second pool. It appears those are side-trips, or data providers, and the actual sequence in pool 1 must be retained by sequence flows. I’m still not sure I know the “why”. It appears to be in support of the next step – conversion to XML or some other approach.

It was Visio 2010 that gave us the issues. I don’t have the diagrams here, but I may send you a copy of a more complex one that illustrates the potential profusion of lines if I enforce the rule to provide incoming and outgoing sequence flows for each activity within a pool.

Thanks again for your patience and information!
Connie

By: Method and Style Validation Update « BPMS Watch

Method and Style Validation Update « BPMS Watch — Sun, 17 Oct 2010 01:28:58 +0000

[...] the rule has gone away. Too bad, as it leads to diagram errors — see Connie’s comment here – so I am going to keep those rules but change to “style [...]

By: bruce

bruce — Sun, 17 Oct 2010 00:49:03 +0000

Connie,
Are you talking about Visio’s built-in Check Diagram or my tool? Maybe this relates to a note I got from Steve White a few days ago re my rule BPMN 0101 and 0102, which says that implicit start and end nodes are not allowed if there are any real start and end events in the process level. That was in fact the rule from OMG in BPMN 1.x and up to the end of FTF in BPMN 2.0. But Steve tells me it got changed at the end of FTF, and so implicit start and end nodes are allowed now.

I told him I would change it to a style rule, because I still think it’s bad practice… and unfortunately, your example kind of proves it (at least as I imagine your diagram — maybe email me a copy). The reason is that without an incoming sequence flow, your activity 4 in pool 1 is an implicit start node of process 1. That means it starts concurrently with activity 1… which it does not. The part of BPMN that is hard for some modelers to catch on to is that a BPMN process is an “orchestration”, a continuous chain of sequence flows from some initial state (start event) to some final state (end event or events). So whatever tool is telling you (mine or Microsoft’s) that you need sequence flow from activity 1 to activity 4 is correct. And it shows why OMG’s caving on this rule was probably a mistake.

By: Connie

Connie — Sun, 17 Oct 2010 00:32:00 +0000

Bruce,
My question stemmed from trying to resolve issues using Visio 2010. Even when we used a message flow from an activity to an event in another pool, we received the error that an activity must have at least one outgoing sequence flow. It seemed that we had to have both a message flow to the other pool and a sequence flow within the first pool. That seems to me to add unnecessary lines and complexity to the diagram. Scenario: activity 1 in pool 1 uses message flow to activity 2 in pool 2, which sequences to activity 3 in pool 2, then message flows up to activity 4 in pool 1. I’m not getting why we need the extra sequence flow from activity 1 to activity 4. (This surely illustrates the need for a diagram instead of words!)
Thanks.

By: bruce

bruce — Sat, 16 Oct 2010 02:20:06 +0000

Connie,
A message flow cannot connect to a gateway. That is just part of the BPMN metamodel, i.e. the structure defined by OMG. A gateway is not a message-aware element. A gateway is just routing logic; it does not take actions like “making” a decision (it can “test” a decision) or sending a message. You can use a message flow to connect an activity in one pool with an activity in another pool. I hope I didn’t say you could not.
–Bruce

By: Connie

Connie — Sat, 16 Oct 2010 01:59:16 +0000

Can you explain the rationale for some of the rules? Specifically, what’s the problem with using a message flow to connect a decision or activity in one pool with an activity in another pool? We finally figured out how to remove the error, but it complicates the diagram. What am I missing?