Yes, that is what is modeled.
The timer starts when Task A is completed. If EITHER Task B or Task C (or both) is not complete when the 45 second timer expires, the not-completed task(s) will be cancelled and activity moves on to Task D.
Though, perhaps I misunderstood since you keep talking about parallel gateways.
If B must be completed before C, you could do:
I believe I expressed myself incorrectly. Based on what you modeled, I understood that in the subprocess, if a task is not solved within 45 seconds, everything in that subprocess stops. However, my intention was that if, for example, task B expires but the previous task was already processed, the process should still continue in the âandâ branch.
Iâm not clear how youâre using the word branch.
the <+> icon indicates âDo these tasks concurrentlyâ not âDo both of these tasksâ
B â C still indicates âDo both of these tasksâ but indicates C will not start before B ends.
Tasks that are completed are not undone unless explicitly told to. So in the Subprocess with the Parallel gateways, B can be done before C, or C can be done before B, but with a Boundary Timer on the overall subprocess, anything not completed in 45 seconds will be told to stop.
Itâs often hard to explain these sorts of things in an abstract way with A, B, C, D rather than real-life examples. The book âReal Life BPMNâ actually helps a lot.
I think Iâm expressing myself incorrectly. After the âand split,â we have tasks B and C. For example, I perform task C, but task B has expired, and in this âand split,â task B disappears, and the âand joinâ sees that there is no task B because it wasnât completed within 45 seconds, and only task C was executed. Should I clarify this better, or show more precisely what I mean?
In the way I drew it, if Task C was completed before the 45 second timer went off, Task B would be cancelled, and the overall process would move on to Task D automatically.
Iâm struggling to come up with a real-world scenario where you would be Ok not doing something that is part of the process if the time limit ran out. If itâs ok to skip it, why is it being done at all?
I keep coming back to the âRegister for Courseâ example, which can be forced to mostly fit:
A = âSet up accountâ
B = âComplete registration formâ
C = âPay Feesâ
D = âAttend Courseâ
But⌠If B is done and C is not, then you wonât be allowed to attend the course.
Ah, I understand your drawing now. What youâre saying makes sense, but my idea was different. I wanted to add a timer event at each task in the parallel gateway. I will create a more detailed model later so that it can be better understood. Essentially, I wanted to try to have one task executed within a specific time frame while the other one is not. However, I believe there might be a problem when merging the gateways.
Excellent! Now weâre working with a concrete example.
Yes, having interrupting timers connecting in to the parallel join will likely cause issues if you try to run it in a BPMN execution engine, but itâs a perfectly valid model for a process description diagram.
You have a redundant boundary timer. To solve this, you can either remove the boundary timers from both âBuy Popcornâ and âBuy Drinksâ or from the sub-process. If you remove it from the subprocess, you can have different durations for âBuy Popcornâ and âBuy Drinksâ, but from a reader perspective, that wouldnât make much sense (Why are you allowed more time to âbuy popcornâ than to âbuy drinksâ?)
If you remove the timer from âBuy Popcornâ and âBuy Drinksâ and leave it on the sub-process, when ârunningâ the process, if the task âBuy Popcornâ is completed within the 45 seconds, and youâre still in line to buy drinks, the timer would go off, you have the popcorn in your hands, you leave the drinks line, and you go sit down to watch your movie.
Is there an alternative way for me to execute it with the engine? My question is, how did you do it above, that you put the timer in the subprocess? Can I still perform the task âBuy Popcornâ but not the task âBuy Drinks,â and then proceed to the task âCheck Ticketâ at the end?
A = Buy Ticket
B = Buy Popcorn
C = Buy Drink
D = Watch Movie
If you buy your drink right away, but get stuck in line for popcorn, you will give up and go watch the movie.
If you buy your popcorn right away, but get stuck in line for drinks, you will give up and go watch the movie.
If you manage to buy both your popcorn and drink in under 45s, youâll proceed directly to watching the movie.
I understand. So, if I make a âbuy drinksâ task but not a âpopcornâ task, can the values I entered in the âbuy drinksâ task appear in a task like âcheck ticketsâ? However, Iâm not sure if the values from the previous task can be seen once the timer runs out.
I think this is where weâre losing clarity in translation.
When I see âIf I makeâ I understand you meaning âput a task in my diagramâ not âComplete the taskâ
As far as I know, yes, variables that have been set on completing âBuy drinkâ (eg. Drink Type: Soda, Drink Brand: Pepsi ) should be available in âWatch movieâ
By âif I make,â I mean to execute the task.
This is a great use of Play Mode to validate both your assumptions and the model. I did a quick recreation of the model @GotnOGuts shared 3 replies above, then executed it in Play Mode. I used the Complete with Variables option on the User Tasks, and assigned a value in B. The value was still available when the subprocess exited.
Thank you both very much. Iâm sorry for asking so many questions.
