Visualization of a reference model

I had a meeting with my mentor Pieter Van Gorp where we talked about the visualization of a reference model as described in previous blog post. We have evaluated the different options and I gave it a second thought, and I want to share those thoughts with you.
Pieter and I agreed to use the option which I call conventional process modeling. This because it is easier to start with and when there is time I could translate this model into the second option (referred to parameterized process modeling).
Because the process models I received from the hospitals show some similarity, a difference analysis and a similarity measurement will be executed, and it is easier to do so on individual process models instead of an integrated (parameterized) process model.
The third option (declarative process model) is less suitable for a reference model, because it does not show a process flow. A process flow shows the most efficient way to execute a process, and when the declarative approach is used this flow is missing. This option is more suitable when the process flow is less strict and the process consists of many exceptions. An approach that tells you what step to do next when you have executed a single task is in that case more suitable.

So for my thesis I will focus on the conventional process models and if there is time these models will be integrated into one parameterized process model. The conventional process model will use simple BPMN2.0 notation (mainly gateways and tasks), while the parameterized model will use the more complex notation (complex gateways, sub-processes, intermediate catching and throwing events, etc., etc.). Which notation is best to use under what circumstances is a different question for a different master thesis project.