Parallel Regions

Some concerns are genuinely independent: a kitchen prepping food while an SLA clock ticks. Modelling them as one flat sequence forces a false ordering. A parallel super state declares two or more orthogonal regions that are all active at once. Each region tracks its own leaf, and the machine’s active configuration is the union of every region’s current leaf.

Open each region with Region, give it an Initial child, and close it with EndRegion. In the food-delivery example, Active runs a Fulfillment region (kitchen then courier) alongside a Watchdog region (an SLA timer):

SuperState(Active).
    Initial(Active).
    Region("Fulfillment").
    Initial(Cooking).
    State(Cooking).On(PlatedUp).GoTo(AwaitingCourier).Assign("recordPrep").
    State(AwaitingCourier).On(PickedUp).GoTo(EnRoute).
    State(EnRoute).
    EndRegion().
    Region("Watchdog").
    Initial(OnTime).
    State(OnTime).After(SLAWindow).On(SLABreached).GoTo(Overdue).Assign("markBreached").
    State(Overdue).
    EndRegion().
    Transition(Active).On(DroppedOff).GoTo(Settling).Assign("recordDrop").
EndSuperState()

Entering Active enters both regions’ initials, so the configuration becomes {Active.Fulfillment.Cooking, Active.Watchdog.OnTime}. An event is broadcast to every region; each region that has a handler reacts independently. A transition on the parent (DroppedOff) exits all regions at once and leaves the whole parallel block.

stateDiagram-v2
    state Active {
        state Fulfillment {
            [*] --> Cooking
            Cooking --> AwaitingCourier
            AwaitingCourier --> EnRoute
        }
        --
        state Watchdog {
            [*] --> OnTime
            OnTime --> Overdue : SLABreached
        }
    }
    Active --> Settling : DroppedOff

The Watchdog is observational: it records a breach without blocking fulfillment, proving regions can advance at their own pace.