Multiparadigm Engine Layer

Documentation status: architecture — see Maturity and evidence.

The multiparadigm-engine layer executes behavior attached to the model without imposing a single computation style. It groups functional, logical, parallel, and neural evaluation paths behind explicit Runtime boundaries.

Principle

The model remains the source of meaning. Choosing an evaluation engine is an execution decision, not a redefinition of the domain.

Responsibilities

• compile or interpret supported executable artifacts;
• create isolated execution contexts;
• orchestrate processes, channels, and tasks;
• move Runtime values with explicit lifetime rules;
• expose errors, cancellation, and diagnostics observably;
• add capabilities without leaking implementation details into the application model.

Maturity

FPL/FPLN and documented reasoning capabilities must be read according to page status and Runtime version. Uncertified neural or Symbolic[Neuro] integrations remain evolution targets and must not be interpreted as available contracts.

Public boundary

Applications use published models, services, SDK/bindings, and Runtime contracts. Internal engine classes are not a public programming surface.