Comments on: A Standard for Case Management?

By: drobert

drobert — Mon, 06 Jul 2009 14:15:29 +0000

I have been working on a Case Management application for a client of mine in the Canadian Federal Government. I have a BPMN background (I have taken one of your courses at a BMPI event, Bruce) so I was keen to model the process using BMPN. As I did, I came upon the limitations of the notation as described above.

I decided to differentiate between “Process Tasks” and “Case Tasks”. Process tasks are, as you would guess, those that involve workflow that either assign the case to various workers or move the case from one phase/state to another. Case tasks are those that involve a single case and can be assigned to and completed by the case team members. They represent the tasks required to complete the case (lamely, are essentially ad-hoc activities).

This architecture was heavily influenced by the fact that my client is developing in an MS Sharepoint environment. Each case will have its own ‘workspace’ with its own list of tasks etc.. but there is no ‘workflow’ at the case level. However Sharepoint provides the possibility for the creation of workflow at the case level as well..

I will have to pick up the 2009 workflow handbook to see how others are tackling this issue.

Dan

By: bruce

bruce — Wed, 01 Jul 2009 00:33:55 +0000

Keith,
I think it should be possible to extend BPMN to better model case management, but I agree with you that OMG is an unlikely source of progress in this area, as the bmi thread clearly indicates. I think we could get further by convening an informal group of interested parties to play around with some ideas around scenarios and notation. I’ll count you in, maybe get Malcolm Ross (Appian) and Paul Tazbaz (Wells Fargo) to contribute as well. All BPMN people. And I’ll trade you a copy of my book for a copy of the 2009 Workflow Handbook.
–Bruce

By: kswenson

kswenson — Tue, 30 Jun 2009 23:42:06 +0000

Good summary. Perhaps this post contains a hint as to why my style of drawing a process diagram represent a “state” which remains static until acted upon by people or other external events. The activities don’t “execute” to completion, but rather are flags that indicate to everyone that the execution is going on, but actually simply wait for the signal that execution is complete. That makes a kind of equivalence between “control flow” and human/external events. Interstage BPM has always offering a continuum of possibilities between completely pre-determined processes and completely dynamic add-each-step-as-you-get-to-it style processes. I am overjoyed to see people discussing the importance of case management. While it is clear that BPMN would need extensions, it remains to be seen if OMG is the place for this discussion.

There are a number of articles on Case Management in the new 2009 Workflow Handbook, especially in regard to use in government.

By: malcolm.ross@appian.com

malcolm.ross@appian.com — Wed, 24 Jun 2009 21:32:02 +0000

Hey Bruce,

I couldn’t agree with you more about the lame support for ad-hoc activities in the current BPMN spec. I think having processes that combine both ad-hoc activities with sequenced activities is much more powerful. As you have also noted, with case management it is important to also look at the relationship of data and process. I have always seen case management processes to be more data centric where processes simply manipulate the data of the case in either sequenced or ad-hoc events. There could be several processes interacting with the same data for a case occuring asynchronously.

Look forward to seeing what recommendations you would offer around case management.

Malcolm Ross
Director Product Management
Appian