What are Exchange Items "good for"?

Hello there,

I am modelling Exchange Items for all Functional Exchanges in my model (in the Physical Architecture layer), giving them classes and units, and establishing data structures.

What can I do with this once I am done?
Do you do this and how/why?

Thank you!

P.S. I looked in Capella Help Contents, in the Github Wiki (metamodel) and in the ARCADIA book by Jean-Luc Voirin for now

The Exchange Items begin to provide a precise definition of the Functional Exchange and the functional dependencies between the two functions. The FE at a basic level has only a name, which if selected appropriately provides a reasonable contextual idea of its purpose. However, to really understand what information one function must produce that another has to consume to create a necessary behaviour requires a data model. Appropriate elements of that data model are than assigned to the Exchange Item. This is the functional interface that eventually one of the realized systems/components will need to implement that others will need to comply to.

1 Like

Dear woodske,

Thanks a lot for the answer!
This bit is for me of particular interest:

How does this happen in Capella?
What do we actually do with the EI, after we fill them with objects from the data model?
Can we somehow reference these variables/classes/interfaces, for example to establish how the receiving component handles them?

Cheers :slight_smile: and thanks again

I think the answer to this question would rather depend on how detailed your model will become and the objective of your modelling. Who will actually implement the function and will they use your model to comprehend the design intent, or will you create a component or system specification that will be given to a third-party to implement in a traditional supplier contractor arrangement.

I do not believe there is a prescribed way to do this either in Arcadia or Capella. At a basic level how a function transforms its inputs into its outputs would be the specification of the function itself. You could write this as a textual description of the function within the model – or leave it as an exercise for the supplier. This along with the relevant elements of the data model can be extracted from the model using something like M2Doc using the EI allocations in the functional exchange to create an interface requirement specification.