Skip to content

Data model

Data Scoping and Visibility

In Collaborative State Machines, data visibility and lifetime are determined strictly by the application's hierarchical containment structure. Data scoping relies on an upward-only resolution mechanism.

A component can access its locally declared variables as well as any variables declared by its direct ancestors up to the root. Conversely, a component can never access variables declared by its siblings or its descendants. This guarantees structural isolation and enforces strict information hiding. For example, a control state can read its parent state machine's data, but a state machine cannot peek into the data of its nested states.

When a variable is referenced, the runtime locates it by searching the accessible scope: it checks the local component first, then traverses upward through the ancestor hierarchy, and finally checks the global persistent data.

Data Categories

To map data lifetime to this structural hierarchy, CSM categorizes data into five distinct classes:

  1. Persistent Data: Declared at the root of the collaborative state machine. This data is globally accessible to all nested components, serving as the application-wide coordination state (though it may incur distributed synchronization overhead).
  2. Transient Data: Declared at the state machine level. This data remains strictly local to that specific state machine instance, bypassing global synchronization and isolating it from sibling state machines.
  3. Static Data: Declared within individual control states. This data is scoped exclusively to the state and persists across repeated entries/activations of that state, maintaining local operational history.
  4. Event Data: Ephemeral payload dictionaries attached to emitted events. These facilitate asynchronous data propagation between state machines.
  5. Service Data: Structured input and output parameters exchanged during external Invoke actions. This forms the explicit boundary between the state machine's internal coordination logic and external computation.

Data Manipulation via Expressions

Data within the accessible scope is manipulated declaratively using expressions. Expressions are utilized to evaluate and update variables (via Eval actions), define conditional progression rules (via Match guards), map dynamic inputs to external services, or populate event payloads.