This question is about how to document process interdependencies rather than solving particular technical BPM problem. We are still learning BPMN and the tools.
We have two processes: Collection and Video production. One of them runs continually, collects text messages from users, and from time to time (usually once a week) it makes a "snapshot" of all newly collected messages (since the last snapshot) and should launch another process to begin some further processing of those messages (sorting, preparation of a video session, transcript, translations, etc.).
Since the first process (collection) effectively runs in an infinite loop, I would model it with a message-throw intermediate event or message-send task. In the second process, I wanted to start with a message start event that is triggered by the message from the Collection process..
I was not able to draw association line between the send task or throw event in Collection and the message start/catch event in the Video production process (assuming they would be in two swimlanes in the same pool, or in two pools in the same BPMN diagram). However, it is not possible to draw association line between such objects, only process flow.
What am I doing wrong? How should I model this type of link between the two processes?