Intent-Oriented Programming Language (IOPL): A Deterministic Foundation for Intent-Centric and AI-Assisted Software Development

Abstract

Modern software systems increasingly demand that developers specify what a system should accomplish rather than how it should be implemented, motivating declarative languages, intent-based domain-specific systems, and AI-assisted programming tools; however, many existing approaches remain confined to narrow domains, are tightly coupled to particular execution engines, or implicitly outsource program meaning to probabilistic models, weakening determinism, reproducibility, and formal reasoning. This paper introduces the Intent-Oriented Programming Language (IOPL), a general-purpose language that elevates intent to a first-class semantic construct while preserving deterministic, verifiable execution: IOPL separates semantic intent from execution strategy via a staged compiler pipeline and a canonical Execution Intermediate Representation (EIR) that freezes observable semantics prior to code generation, and it constrains AI components to non-authoritative roles whose outputs must conform to EIR and be subject to verification. We outline the language model, compiler architecture, runtime semantics, determinism contract, and verification approach, and situate IOPL within the broader research landscape.