Right now there is an assumed 1:1 mapping between a task and its execution. Some "task types" (e.g., an environment definition) will map to multiple executable actions. We should refactor our abstraction into two parts: a logical task graph and a physical execution graph.
Right now there is an assumed 1:1 mapping between a task and its execution. Some "task types" (e.g., an environment definition) will map to multiple executable actions. We should refactor our abstraction into two parts: a logical task graph and a physical execution graph.