Types for Grassroots Logic Programs

2026-01-25

AI Summary

Existing Grassroots Logic Programs (GLPs) lack type support for interactive, potentially deadlocking, or non-terminating computations, making it difficult to ensure communication reliability in AI-assisted development. This work proposes the first modal path regular type system incorporating directional modes, which distinguishes between read and write directions of subterms to enable static type checking for partial, interactive, and even non-terminating programs. By integrating covariance and contravariance semantic conditions with formal syntactic rules, the system establishes that well-typedness is equivalent to semantic paths satisfying directional constraints. Furthermore, a correct-by-construction Dart implementation is automatically generated from the formal specification by an AI, demonstrating the feasibility of human-AI collaborative programming.

Technology Category

• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Knowledge Representation and Reasoning: Logic Programming

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Grassroots Logic Programs (GLP) is a concurrent logic programming language in which logic variables are partitioned into paired readers and writers. An assignment is produced at most once via a writer and consumed at most once via its paired reader, and may contain additional readers and/or writers. This enables the concise expression of rich multidirectional communication modalities. ``Logic Programs as Types for Logic Programs''(LICS'91) defined types as regular sets of paths over derivable ground atoms. Here, we define types to be regular sets of moded paths, where a mode captures directionality of communication -- whether a subterm is consumed from or produced to the environment -- enabling the typing of interactive partial computations including those that eventually deadlock or fail, or never terminate. We provide a syntactic definition of well-typing and prove that a program is well-typed iff the path abstraction of its moded-atom semantics satisfies covariance and contravariance conditions with respect to its type. The GLP type system was implemented in Dart by AI, starting from a mathematical specification of Typed GLP (this paper), deriving from it an English spec (written by AI), and from the spec deriving Dart code (by AI). While GLP is naturally untyped, the motivation for Typed GLP comes from programming with AI: Asking AI to program complex communication modalities in GLP (and in general) and hoping for the best is a tenuous strategy. The emerging discipline we advocate and employ is for the human designer and AI to jointly develop and agree upon (1)~GLP types; (2)~GLP procedure type declarations; (3)~informal (English) descriptions of the procedures; and only then let AI attempt to write (4)~GLP code based on those.

Problem

• Grassroots Logic Programs
• type system
• concurrent logic programming
• interactive computation
• AI-assisted programming

Innovation

• moded paths
• type system
• concurrent logic programming
• covariance and contravariance
• human-AI collaboration