Comments on: My New BPMN Wish List http://www.brsilver.com/wordpress/2008/07/17/my-new-bpmn-wish-list/ Bruce Silver's blog on business process management Wed, 07 Jan 2009 19:01:12 +0000 http://wordpress.org/?v=2.7 hourly 1 By: conrad http://www.brsilver.com/wordpress/2008/07/17/my-new-bpmn-wish-list/comment-page-1/#comment-6866 conrad Thu, 09 Oct 2008 15:12:05 +0000 http://www.brsilver.com/wordpress/?p=355#comment-6866 Bruce, Here's a draft of BPMN-S/BPDM that linking notation and metamodel: http://doc.omg.org/bmi/08-09-07 (browse the metamodel at ftp://ftp.omg.org/pub/docs/bmi/08-09-05/pages/188c21b53f42002f.htm) See Sections 7.6 (Process Model) and 7.9 (Interaction Protocol). These sections can be read standalone from the rest of the spec. See Figure 40 (Events and Successions) regarding temporal constraints (these need to be updated to the latest IOS notation for non-interrupting events). Conrad Bruce,

Here’s a draft of BPMN-S/BPDM that linking notation and metamodel:

http://doc.omg.org/bmi/08-09-07
(browse the metamodel at ftp://ftp.omg.org/pub/docs/bmi.....42002f.htm)

See Sections 7.6 (Process Model) and 7.9 (Interaction
Protocol). These sections can be read standalone from the
rest of the spec. See Figure 40 (Events and Successions)
regarding temporal constraints (these need to be updated to
the latest IOS notation for non-interrupting events).

Conrad

]]> By: Event Handling in BPMN 2.0 - BPMS Watch http://www.brsilver.com/wordpress/2008/07/17/my-new-bpmn-wish-list/comment-page-1/#comment-6837 Event Handling in BPMN 2.0 - BPMS Watch Wed, 17 Sep 2008 17:20:23 +0000 http://www.brsilver.com/wordpress/?p=355#comment-6837 [...] obviously a work in progress with some very rough edges.  One major enhancement is my favorite wish list item, the non-aborting attached event.  The proposal provides this for timer, message, conditional [...] [...] obviously a work in progress with some very rough edges.  One major enhancement is my favorite wish list item, the non-aborting attached event.  The proposal provides this for timer, message, conditional [...]

]]> By: sca http://www.brsilver.com/wordpress/2008/07/17/my-new-bpmn-wish-list/comment-page-1/#comment-6829 sca Wed, 06 Aug 2008 11:44:30 +0000 http://www.brsilver.com/wordpress/?p=355#comment-6829 The wishes for variations of "non-aborting attached events" reminds me of my Master thesis work in 1985 with something originally called PiP (Processes interfaced through ports), but that later was known as PPM (Process Port Model from Arne Sølvbergs IS group at the Norwegian Institute of Technology). We had the concept of a "free flow", which essentially was a kind of sequence flow that could be sent without the activity terminating. A "free sequence flow" variant in BPMN would solve several of these wishes. In particular, it also would be useful to combine with conditional sequence flow to indicate things that "may happen", or be triggered from the mother activity while the mother is still alive. Bruce's mentioning of temporal logic reminds me about Allen's temporal interval relations which lists the 7 possible relationships between two activities time intervals; for an example see http://www.dcs.kcl.ac.uk/staff/endriss/cs3aur-02/pdf/time.pdf Steinar The wishes for variations of “non-aborting attached events” reminds me of my Master thesis work in 1985 with something originally called PiP (Processes interfaced through ports), but that later was known as PPM (Process Port Model from Arne Sølvbergs IS group at the Norwegian Institute of Technology). We had the concept of a “free flow”, which essentially was a kind of sequence flow that could be sent without the activity terminating. A “free sequence flow” variant in BPMN would solve several of these wishes. In particular, it also would be useful to combine with conditional sequence flow to indicate things that “may happen”, or be triggered from the mother activity while the mother is still alive.

Bruce’s mentioning of temporal logic reminds me about Allen’s temporal interval relations which lists the 7 possible relationships between two activities time intervals; for an example see http://www.dcs.kcl.ac.uk/staff.....f/time.pdf

Steinar

]]> By: bruce http://www.brsilver.com/wordpress/2008/07/17/my-new-bpmn-wish-list/comment-page-1/#comment-6823 bruce Mon, 28 Jul 2008 15:13:30 +0000 http://www.brsilver.com/wordpress/?p=355#comment-6823 Conrad, If your goal was to get me to go back and re-read the BPDM submission, you've probably accomplished it. My recollection is the notation part was particularly unrevealing, so I may come back to ask you for page references, or even to say it in plain English. --Bruce Conrad,
If your goal was to get me to go back and re-read the BPDM submission, you’ve probably accomplished it. My recollection is the notation part was particularly unrevealing, so I may come back to ask you for page references, or even to say it in plain English.
–Bruce

]]> By: conrad http://www.brsilver.com/wordpress/2008/07/17/my-new-bpmn-wish-list/comment-page-1/#comment-6822 conrad Mon, 28 Jul 2008 13:27:14 +0000 http://www.brsilver.com/wordpress/?p=355#comment-6822 Bruce, Good ideas as usual. Most of them are supported in standard BPDM, including notation (normative in the BPDM-based BPMN2 submission). Temporal constraints are the foundation of standard BPDM and BPDM-based BPMN 2, while the IOS BPMN 2 uses UML-style token flow and cannot support the level of expressiveness you're looking for. See http://www.conradbock.org/#BPDM for more about BPDM. > A second key point: BPMN is really just about when > activities occur in relation to each other, when they > start, when they end, and what happens next. It doesn’t > have a lot to say about how (the data, the > implementation), and just scratches the surface of who > (pools and lanes). BPMN has some data flow capabilities, if you include the attributes. Standard BPDM and BPDM-based BPMN 2 has a full data flow model. IOS BPMN 2 is still working on theirs. > But there is ambiguity about what is meant by ’starting.’ > In BPMN an activity starts when it is enabled by incoming > sequence flow. But to a business user, it starts when > someone or something begins working on it. > The time when a task is assigned and made available in a > performer worklist is not the same as the time when the > performer begins working on it. This distinction is > fundamental to human workflow, but ignored by BPMN. All the BPMN variants (standard BPMN, standard BPDM, BPDM-based BPMN 2, and IOS BPMN 2) distinguish the start of a process from the start of the first step in the process. Perhaps the first step in the process can be taken as "someone or something begins working on it". It could be notated as attached to an activity with a new event symbol referring to the first step. > Wish item #1 is the non-aborting attached event. There > is consensus on the value of the semantic; we just need a > notation. I propose to use the start event notation > (single thin circle) drawn attached to an activity for > this purpose. The thin circle currently means "start" (before work actually begins) and I think modelers would assume it meant the start of the activity it is attached to, analogous to the error attached event already in BPMN. I'd reserve this notation for the SF/FS/SS semantic below. Perhaps the double border on any attached event could indicate it is non-aborting. > It means that the activity (task or subprocess) cannot > complete until the event occurs. In standard BPDM and BPDM-based BPMN 2 this would be an attached end event with an incoming sequence flow. This establishes a temporal constraint that the activity cannot finish before the source of the sequence flow does. IOS BPMN 2 cannot support this with it's current UML-style token semantics, see below. > The Temporal Perspective: Expressing Temporal Constraints > and Dependencies in Process Models can be found in the > 2008 BPM and Workflow Handbook published by WfMC Thanks for the citation, this is exactly the BPDM approach to process semantics (for example, see slide 4 and after in http://www.conradbock.org/bock-bpmn-2-business-process-semantics-web.pdf). The IOS BPMN 2 submission, by contrast uses UML-style token flow, which is not expressive enough to capture these constraints, is too computational from a business perspective, and doesn't support choreography semantics. > In a semi-structured process, such as described by a tool > like Microsoft Project, FS is only one of several > dependencies between activities. You also have: > Start-Start (SS). > Finish-Finish (FF). > Start-Finish (SF). > Finish-Start(FS) BPDM can model all these, because BPDM sequence flows ("successions") can specify an event on each end of the sequence flow. The notation (nonnormative in BPDM, normative in the BPDM-based BPMN 2 submission) is similar to yours, using the existing event symbols (and the additional ones for success and failure) attached to activities. IOS does not support these in either notation or semantics, and in semantic has a significant barrier due to its UML-style token semantics. Conrad Bruce,

Good ideas as usual. Most of them are supported in standard BPDM, including notation (normative in the BPDM-based BPMN2 submission). Temporal constraints are the foundation of standard BPDM and BPDM-based BPMN 2, while the IOS BPMN 2 uses UML-style token flow and cannot support the level of expressiveness you’re looking for. See http://www.conradbock.org/#BPDM for more about BPDM.

> A second key point: BPMN is really just about when
> activities occur in relation to each other, when they
> start, when they end, and what happens next. It doesn’t
> have a lot to say about how (the data, the
> implementation), and just scratches the surface of who
> (pools and lanes).

BPMN has some data flow capabilities, if you include the attributes. Standard BPDM and BPDM-based BPMN 2 has a full data flow model. IOS BPMN 2 is still working on theirs.

> But there is ambiguity about what is meant by ’starting.’
> In BPMN an activity starts when it is enabled by incoming
> sequence flow. But to a business user, it starts when
> someone or something begins working on it.

> The time when a task is assigned and made available in a
> performer worklist is not the same as the time when the
> performer begins working on it. This distinction is
> fundamental to human workflow, but ignored by BPMN.

All the BPMN variants (standard BPMN, standard BPDM, BPDM-based BPMN 2, and IOS BPMN 2) distinguish the start of a process from the start of the first step in the process. Perhaps the first step in the process can be taken as “someone or something begins working on it”. It could be notated as attached to an activity with a new event symbol referring to the first step.

> Wish item #1 is the non-aborting attached event. There
> is consensus on the value of the semantic; we just need a
> notation. I propose to use the start event notation
> (single thin circle) drawn attached to an activity for
> this purpose.

The thin circle currently means “start” (before work actually begins) and I think modelers would assume it meant the start of the activity it is attached to, analogous to the error attached event already in BPMN. I’d reserve this notation for the SF/FS/SS semantic below. Perhaps the double border on any attached event could indicate it is non-aborting.

> It means that the activity (task or subprocess) cannot
> complete until the event occurs.

In standard BPDM and BPDM-based BPMN 2 this would be an attached end event with an incoming sequence flow. This establishes a temporal constraint that the activity cannot finish before the source of the sequence flow does. IOS BPMN 2 cannot support this with it’s current UML-style token semantics, see below.

> The Temporal Perspective: Expressing Temporal Constraints
> and Dependencies in Process Models can be found in the
> 2008 BPM and Workflow Handbook published by WfMC

Thanks for the citation, this is exactly the BPDM approach to process semantics (for example, see slide 4 and after in http://www.conradbock.org/bock.....cs-web.pdf). The IOS BPMN 2 submission, by contrast uses UML-style token flow, which is not expressive enough to capture these constraints, is too computational from a business perspective, and doesn’t support choreography semantics.

> In a semi-structured process, such as described by a tool
> like Microsoft Project, FS is only one of several
> dependencies between activities. You also have:

> Start-Start (SS).
> Finish-Finish (FF).
> Start-Finish (SF).
> Finish-Start(FS)

BPDM can model all these, because BPDM sequence flows (”successions”) can specify an event on each end of the sequence flow. The notation (nonnormative in BPDM, normative in the BPDM-based BPMN 2 submission) is similar to yours, using the existing event symbols (and the additional ones for success and failure) attached to activities. IOS does not support these in either notation or semantics, and in semantic has a significant barrier due to its UML-style token semantics.

Conrad

]]> By: bruce http://www.brsilver.com/wordpress/2008/07/17/my-new-bpmn-wish-list/comment-page-1/#comment-6811 bruce Mon, 21 Jul 2008 03:42:13 +0000 http://www.brsilver.com/wordpress/?p=355#comment-6811 Scott, You make a good point. A number of BPMS offerings capture the 'started' event for BAM but any triggered process actions are limited to BAM capabilities and are 'off the map' as far as the process model is concerned. The BPMN standard definitely has pushed BPMS vendors to add more event-triggered functionality, so what you suggest is not out of the question. Scott,
You make a good point. A number of BPMS offerings capture the ’started’ event for BAM but any triggered process actions are limited to BAM capabilities and are ‘off the map’ as far as the process model is concerned. The BPMN standard definitely has pushed BPMS vendors to add more event-triggered functionality, so what you suggest is not out of the question.

]]> By: sfrancis http://www.brsilver.com/wordpress/2008/07/17/my-new-bpmn-wish-list/comment-page-1/#comment-6810 sfrancis Mon, 21 Jul 2008 01:25:14 +0000 http://www.brsilver.com/wordpress/?p=355#comment-6810 Bruce - I couldn't agree more that these notions are not currently well-captured in BPMN. They're also not well implemented by vendors, and I think this is a case where an improvement in the spec would be followed by improvements in the implementations... However, it is also an opportunity for Vendors to augment their implementations to anticipate this need (regardless of future notation, these kinds of time-related configuration details could be captured in "properties" below the visual level). Capturing these details below the visual level is hardly ideal, as it means you can't see the semantics of the ordering without inspecting the implementation details, but it at least prepares the way for vendors to support a more explicit diagrammatic approach... I would say there are a lot of scenarios where you really want to know "start" as the moment a user starts the activity. So, a time-based escalation might not need to happen if a user has already "started" the activity, but should happen if no user has started it yet... etc. (like you, i don't care what the specific diagram choices are - I'm not iconographer :) scott Bruce - I couldn’t agree more that these notions are not currently well-captured in BPMN. They’re also not well implemented by vendors, and I think this is a case where an improvement in the spec would be followed by improvements in the implementations… However, it is also an opportunity for Vendors to augment their implementations to anticipate this need (regardless of future notation, these kinds of time-related configuration details could be captured in “properties” below the visual level). Capturing these details below the visual level is hardly ideal, as it means you can’t see the semantics of the ordering without inspecting the implementation details, but it at least prepares the way for vendors to support a more explicit diagrammatic approach…

I would say there are a lot of scenarios where you really want to know “start” as the moment a user starts the activity. So, a time-based escalation might not need to happen if a user has already “started” the activity, but should happen if no user has started it yet… etc.

(like you, i don’t care what the specific diagram choices are - I’m not iconographer :)

scott

]]>