RM
Often in modeling business processes, in any tool, there is the issue of assuring that all the diagrams are well conected among each other.
I would like to know how is the best (or at least less manual) way to do so?
Let me explain it with apples... in theory, this should happen in every connection...
- in diagram A- output from A to B with an event XX
- in diagram B- input form A to B with an event XX
How to be sure that this is happening every time... avoiding that some part of the equation is missing, that can be the output, the input or that the event is a different object like XX - Xx ... ???
I'll be PLEASED to hear some comments. Thanks.
You are touching a difficult point. Of course there are situations where an automatic maintenance of the diagrams to keep consistency would be desirable. But there are also many cases where this behavior would be wrong.
The real model in ARIS consists of the object definitions, the connetions between them and their attributes. The diagrams only create special view on this object net. So customer is free to decide, what part of the underlying object net should be visible in a certain context. If now the tool automatically puts elements in a diagram this might lead to not wished results.
I could imagine to implement macros that are activated everytime a connection between objects has been changed and that keeps consistency with other diagrams to fulfill your requirement but would not make it a standard funcionality, because the user expectations are too different at this point.
The situation with events is just one of many formal rules you usually follow when modeling processes, organizational charts, application systems, etc. For example, if you connect the EPC models with “process interface” objects, you will also have to check their consistency. And this is part of the modeling methodology you follow in your company. I agree with Uwe that automatic maintenance of object occurrences and connections is not very wise.
ARIS Business Architect provides quite nice “semantic check” that allows you to check your models and object with lots of formal rules. You can also create a script of your own that will check your models with the formal rules that are important for you and provide you a report. E.g. you can check all EPCs and see if a branch ends with a “process interface” object, whether the preceding “event” object has occurrence in the next process. Well, this is "a less manual way" to assure quality of your models.
I don't wish that the system make an auto-completion of the diagrams, that would be a huge mistake.
I just like to spot the dead ends, so the people responsible of the different areas agree on whether or not there is a link between two processes; and if it is also to agree on the product received and its characteristics.
I think I understand your problem. The usage of process interfaces should be sufficient to reach the goal of graphical marking of process starts and ends as connected with other processes. If you link your processes directly by events the navigation pane supports you to find occurences in other processes. And you could introduce specific event symbols that show the linkage character. And with macros you can check and assure the consistency of your modeling.
