Errata of Kunze, Weske: Behavioural Models
If you find errors in the book or if anything is written in a way that is unclear to you, please drop me a note at mathias dot weske at hpi dot de. I will try to clarify things in the next edition; all known errors will be posted on this web page.
- In Figure 5.27 on page 149, transition t1 needs to be deleted, because the subprocess can only be cancelled after it has started. It only starts after the Get Paper transition has begun.
- For the same reason, in Figure 5.29 on page 150 the Get Paper begin transition puts a token on p and not the Accept Reviewing transition.
- In Figure Figure 6.9, the state space is not correct, since the value of the state id is not taken into account.
- We came up with a formalization of the 'multiple instances without a priori runtime knowledge' control flow pattern, using a coloured Petri net. While instances of B can be started, a transition can add tokens to the input place of B. Therefore, at the start of B, the number of instances does not need to be known. To make sure C synchronizes correctly, all additional tokens have to be removed first. This is represented by the inhibitor arc in the figure below. C can fire if there are m tokens left in its input place.