pulltypes

Pulltypes are a family of type-theoretic constructs designed to model and enforce pull-based interactions in programming languages. In a pulltype system, data producers do not initiate communication; instead, consumers issue explicit requests for data, and the type system enforces that every data item is fetched only after a corresponding request. This contrasts with push-based or event-driven models, where producers emit data without awaiting consumer readiness.

The concept arises from dataflow and streaming debates and is often discussed in connection with session types, lazy evaluation, and resource-aware type systems. A pulltype of a value carries not only the value's type but also protocol information describing how the value is requested, supplied, and possibly released. Types can encode sequences of requests, backpressure constraints, or align with usage patterns to prevent resource leaks and ensure correct ordering of operations.

In practice, a language with pulltypes would allow functions to declare that they return a value only when a caller pulls, and the type checker would guarantee that all required requests happen in a well-formed order. This can enable safer streaming pipelines, demand-driven I/O, and efficient memory management, particularly in resource-constrained or real-time settings. Pulltypes can interact with effects and mutability, requiring careful design to avoid circular dependencies or deadlock.

Criticism centers on added complexity and the absence of a standard approach; supporters argue that pulltypes provide fine-grained control over resource usage and enable predictable performance characteristics. The term remains primarily used in theoretical discussions and niche language implementations.