How can optional inputs to a process and used by a system function be modeled?

Interesting but please clarify the case you want to solve. What happens if none or both are available? Will the system function wait until at least one data item is available?

In any case you may consider the following solution-candidate:

Hi,

I would vote for two uniquely named System Functions. Typically, the logic applied in David's example goes something like this: "It's exactly the same function, it just uses different data." The problem here is that you can't use the words "same" and "different" in the same sentence. If you do, essentially what you're saying is that it's "exactly the same....except for where it isn't." IMHO, in order to re-use a function object all objects allocated to the function (e.g. objects that would be on a FAD), have to be the same. So, if it's the same system, using the same data, stored in the same datastore, using the same transaction code, with the same avg. processing time, the same KPI, in the same location, etc. then you can re-use it. If all these things are NOT the same, then it's simply different and should be so indicated. Probably more relevent when using Simulation and or creating test cases. Just my opinion, of course.

Best Regards,

Bob

Thanks for the advice Ivo and Robert.  In the case I am working on, only one or the other cluster will be present.  (There is no chance that none or both will arrive as input.)  The clusters represent an input to start a process, input to a Web Service.  The process isn't waiting.  

Thanks for your example Ivo, I have used your example in the model.  I like the notion of an event having state by associating a cluster with it.

 Robert, I appreciate your point of view however I was thinking along the lines that the function could be overloaded.  i.e. different combinations of parameters/inputs are accepted and affect how or what the function responds/does.  

Thanks again.

David Kerins