The language is case-sensitive. Keywords are reserved and cannot be used as identifiers. The primary reserved keywords in v10b are:
let, inductive, rule, proof, term, type, forall, exists, :=.
Today I’m excited to announce the finished release of FILF Version 10b — a focused, stable, and feature-complete update that polishes core workflows, tightens performance, and introduces a few highly requested refinements. Below is a concise overview of what changed, why it matters, and how to get the most from the update.
FILF was initially conceived as a tool for verifying properties of distributed systems. Early versions relied on a Hindley-Milner type inference engine. However, as the complexity of the embedded logics grew, it became apparent that simple type inference was insufficient for dependent types. filf finished version 10b
Version 10b moves away from pure Hindley-Milner, adopting a Stratified Dependent Type System (SDTS). This shift allows for type-level computation without the undecidability issues that plagued Version 9.
The development of Functional Inductive Logic Frameworks (FILF) represents a convergence of functional programming paradigms and automated theorem proving. For the past decade, the FILF project has sought to create a meta-logic capable of embedding complex domain-specific logics without the overhead traditionally associated with proof assistants. The language is case-sensitive
The release of FILF Version 10b marks the "finished" milestone in the 10-series lifecycle. Unlike the experimental releases (v10a0 through v10a5), Version 10b is characterized by a frozen kernel and a standardized standard library. The primary motivation for this version was the resolution of the Recursive Descent Paradox identified in late Version 8, which limited the framework's ability to handle self-referential inductive definitions.
This paper serves as the canonical reference for the v10b specification. Section 2 outlines the necessary background; Section 3 details the syntax; Section 4 covers the type theory; Section 5 presents the implementation details; and Section 6 evaluates the performance. This prevents the construction of types that contain
The core contribution of FILF v10b is the finalization of the Stratified Dependent Type System.
To ensure decidability of type checking, v10b enforces a hierarchy of universes:
This prevents the construction of types that contain themselves, eliminating the "Type-in-Type" inconsistency.