European BPM Round Table

It wasn't trending topic (#EuropeanBPMRoundTable) on twitter, unfortunately, but it was very interesting though. The BPM Round Table isn’t something medieval, and BPM has nothing to with music, at least not in this definition. It is a discussion group with representatives from both research and companies, and they present their work which has something to do with Business Process Management (BPM).

The presentations confirmed the subject and relevance of my master thesis project and it is nice to see that the work I am doing can be helpful for some researchers, or that they ran into similar problems. The presentations from the BPM Round Table provided me with some new insights, which I would like to present below and ask you to give your opinion. It is just a brainwave which I gain, inspired by some of the talks today so please be critical. Also please spread the word, because I need some input for this idea.

My thesis is about reference modeling in healthcare and about evaluating BPMN as an appropriate tool for this reference modeling, as you all might know (and if not CLICK HERE). These reference models will eventually be the bases for new process models, but I will not be modeling those. In a reference model you are able to choose between paths that best fits a specific organization, but how am I going to model these different paths? In my opinion there are three ways to model this.

First is the conventional way by modeling all possibilities for a specific organization. These possibilities can be captured within a sub-process, but all traces have to be included.

The second way is to parameterize a process model. So one process model is shown with all possible paths included, and the parameters select which path to follow.

The third way is to use the declarative approach. In this model I only specify which sequence of tasks is not allowed, which means that the rest is allowed. This means I don’t have to model each possibility
for each organization, which can have a huge impact on the modeling process in terms of time.

I illustrated the three possible ways to model by using an example. The example uses the morning ritual process most of the people follow each morning after waking up. Some people start their day by taking a shower, other by taking a bath and the rest might not take a shower at all in the morning. Some people eat breakfast and make lunch, some only eat breakfast and some don’t do any of these two. I assumed in this example that everybody dresses before they leave and brushes their teeth.

If I use the conventional way to model the process described in the example above I have to model all possibilities. In figure 1 you see only three possibilities, but lots more are possible.

Figure 1 - Conventional process model

If I use parameters to show the process from the example the notation becomes more complex, but there is only one process model needed. The parameters will tell you which sequence to follow. So one person takes a shower and does not take breakfast neither makes lunch, while another person does not take a shower, but makes lunch while he is eating his breakfast. See figure 2 for the illustration.

Figure 2 - Parameterized process model

Figure 3 shows the declarative approach. After you ate breakfast you brush your teeth, but you can do this after you took a shower and after you made lunch. In this case I don’t have to model at all, but I can define the constraints and as long the constraints are met, any sequence of process steps is possible.

Figure 3 - Declarative process model

Please let me know by leaving a comment how you think about the ideas presented above and which idea is best to visualize a reference model. Thanks in advance!

(You can enlarge the images by clicking on them)