Documenting the Big Picture

[tony markos]

The big picture of a system is characterized as having processes that can occur in any order and/or processes that can occur at any time.  In short, there is no definedable sequence for processes at this level.   What technique is used to capture the big picture?  (Note: BPMN is a sequence oriented technique.)

Thanks,

Tony Markos

[Thelonius]

Sounds like a State Machine.

[KP]

Sounds like homework.

[fwoolz]

A wide-angle lens?

[Dorian Workman]

Multiple, unrelated processes.  I prefer Activity Diagrams to document processes.

[Geert Bellekens]

Sounds more like an exam question to me

If I were to get a question I would respond with a couple of counterquestions:
- This "System", what it meant by that? Is it a computer system, a business, a company, or an eco-system, or... The type of system might influence the way it is modelled.
- Same goes for these processes, are these automated processes, business processes,...
- And then for the technique; there are a lot of techniques that could be used, from different methodologies, ranging from free text documents to diagrams in UML, BPMN, SYSML,...

So there is no single definite answers to a question that open.

Geert

[son-of-sargasso]

Genesis I:1 "In the beginning..."

But given that there seems to be no expressed reason for modeling this system in the first place then I'd just go with the Heissenberg Uncertainty Principal - "any system will be perturbed by any measurement placed upon it" (or several million words to that effect - but please... DONT START!   )

bruce

[salayande]

The big picture depends on what you consider to be a system. A system could be a software , enterprise or organizations collaborating to serve a common customer as in an airport. In these cases, the big picture suggests a blueprint (application, enterprise or ecology blueprints) in which case you have a number of choices depending on the view you are required to show your target audience. If the view of interest is the functional decomposition of an ERP (as in the BaaN IV and SAP Blueprints), get a copy of the Zachman Add-In and overlay the functions over each other to create the functional blueprint. Given that a business function is a non-sequential view of a business capability, the blueprint should address your concern.

kind regards

Segun

[Andrew Warner]

Hi all,
I don't think there is a single big picture. As you say, the system is so complex there are many different characteristics that might be of interest to someone at some time.

Since a model is a pale imitation of reality it is clearly to be used for some purpose.

If you want a strategic overview, the business motivation model http://www.omg.org/technology/documents/br_pm_spec_catalog.htm, which is part of the OMG's Business Strategy, Business Rules and Business Process Management Specifications, might help.

For a process view Alec Sharp's approach to process analysis and design is fabulous. See http://www.clariteq.com/. He has also written a wonderful book, Workflow Modeling, which is well worth the price.

If you want an overall system veiw you could look at the viable system model. It's a model of any viable system that provides a structure that must be present in order for the system to be viable. It's a good platform to try to understand what mechanisms are or are not in place and what structures are making the system more or less viable. It doesn't remove the necessity to understand all the inter-related disciplines but it does allow you to focus where it's really important. You can find some stuff here http://www.managementkybernetik.com/en/fs_methmod3.html but just search for "viable system model" and you'll get plenty.

The other thing is that there is probably a limit to how much we can model in a deterministic way and still get use out of it. Once we get into complexity where cause and effect are not so obviously related we need other ways to deal with it. The cognitive edge group has an approach they call cynefin. There's a quick intro here http://www.youtube.com/watch?v=5mqNcs8mp74.

I hope that helps.
Cheers,
Andrew.

[tony markos]

Segun:

Thanks for your suggestion.  The Zachman Framework recommends process modeling technques for the high-level modeling of systems from a buisness perspective. (By "system" I mean a either a manual, automated, or combination manual/automated system.)  But that recommendation is incorrect as process modeling techniques are sequential in nature, whereas systems at the high level have no defined sequence.  One can not model a non-sequential system using a sequence-dependent technique.

Geert:

Thanks for the response.   However, I don't see what difference whether the system consists of people, computers, or both makes?     If capturing the big picture depends on making distinctions of such, how is one to capture an integrated big-picture for an enterprise-wide system who's processes are implemented by a combination of both software and people?  

In additon to evaluating the BPMN in terms of the question that I have,  I have researched the functional/process modeling technqiues associated the UML and, per your suggestion, the SYSML.  The only possible technique for handling processes that can occur in any order are Use Cases.  

However, Use Cases are not a "big-picture", enterprise-wide modeling technique.   They for example are not used to capture totally manual processes.

(BTY:  My question is neither for an exam or for a homework assignment.   I am "in the field")

Guys, if any other recommendation comes to mind, please let me know!

Thanks,

Tony Markos

[KP]

In additon to evaluating the BPMN in terms of the question that I have,  I have researched the functional/process modeling technqiues associated the UML and, per your suggestion, the SYSML.  The only possible technique for handling processes that can occur in any order are Use Cases.

How about BPMN's ad-hoc processes?

EDIT: In fact, how about a UML activity diagram without control flows? I think this was what Dorian suggested higher up the page.

[Geert Bellekens]

However, Use Cases are not a "big-picture", enterprise-wide modeling technique.   They for example are not used to capture totally manual processes.

Tony,

I wouldn't be so sure about that. There is the concept of "Business Use Cases" that are used in a Computational Independent way (CIM level).

All I'm saying is that there are a lot of techniques out there, and one might be more appropriate then the other depending on the actual domain and abstraction level.

Geert

[tony markos]

All:

Thanks for the additional suggestions!  Got some more research to do.

Tony Markos

[paddler]

Hi Guys

 I have just had to document business processes at a very high level to look for means of automation.

 I settled on something mentioned in "Practical Software Engineering" by Enricos Manassis. He advocated using business use cases too describe ALL steps in a process (manual and system), identifying the business actors/workers and then drawing activity diagrams that mapped to the Use Cases / Actors ( each actor in the Use Case gets a swimlane)

his book: http://cgi.ebay.com/Practical-Software-Engineering-:-Enricos-Manassis-(Paperback,-2003)_W0QQitemZ341308425966QQcmdZViewItemQQimsxZ20091207?IMSfp=TL091207207010r7825

It has worked well for me as I can go back to my UML "as is" models and copy/paste them into new diagrams and update them based on user feedback.

 So, Business Use cases and Activity Diagrams get a thumbs up from me

PT

[BillyMac]

I would really love to get acces to this solution as well.
I have emailed Dorian but so far no response.
Does anyone have any additional help regarding HTML report searching solutions?