MB

Hi all

We have different ways of modeling in Aris, depending on which viewpoint is used to model, or who is modeling.

Let's suppose project A is mainly focused on business architecture.

So we're using these models :

  • BPMN diagrams with activities and data objects as input or output of these activities
  • Matrix models to record the relationships between an activity and applications that supports the activity . (this is the case when the application provides the data to the activity or stores the data output of the activity, or when the activity is carried out using the application)

In this case we want to be albe to infer for example: 

  • if activity A1 is carried out using  application A2  (relationship is recorded in Aris model)
  • if A1 has data D has output (relationship is recorded in Aris model)
  • then application A2 creates data D (relationship is not recorded in Aris model, this is a "virtual" derived or implicit relationship) 

In another project B, we're more focusing on the application layer, and we're modeling the interactions between applications and data (that is we design models where applications access (read or write) data , the relationship is explicit and stored in Aris)

Now we want to design a query or a report that outputs all applications and data used by those applications .

The question then is : how to get all the relationships between data and applications  ? (In our example above we want to get the relationship between application A2 and data D).

I see some options, but none seems entirely satisfactory :

  1. Allways explicitely record all relationships in models.  The queries/reports rely only on existing relationships (more work when modeling, less when writing queries)
  2. Define rules for implicit relationships derivation , theses rules must be taken into account when writing queries (less work when modeling, more when writing queries)

Thanks for your inputs