hermes-brain/concepts/cl-modernization-initiative.org

CL Modernization Initiative via Passepartout

• The Initiative
• Four Phases
• Dependencies
• Key Principles

The Initiative

A plan to bring Common Lisp from the 1990s to modern standards using Passepartout as an autonomous software engineer. The gap is ~25 years of deferred ecosystem work. An agent working at 10x human velocity closes it in 2-3 years.

Four Phases

1. Build System — cl build / cl test / cl doc that works like Cargo. Lockfiles, single namespace, integrated test/doc runner, parallel compilation. Wraps ASDF with modern conventions via a CL.toml manifest.
2. Tooling — LSP server that surfaces SBCL's world-class type inference. Jump-to-definition, type-on-hover, rename, refactoring. Online (connected image) and offline (static analysis) modes. Integrates with Coalton for typed paths.
3. Language Modernization — Coalton as first-class opt-in typing. Modern standard library (hash sets, priority queues, JSON, HTTP, async, immutable structures). Unicode-by-default. Pattern matching. All backward-compatible with the 1994 standard.
4. Prover Bootstrap + Deployment — HOL kernel in CL verified by ACL2. Screamer-backed proof search. HOL proves meta-level properties (macro soundness, evaluator equivalence, concurrent safety). Static binary deployment via cl build --release.

Dependencies

Phase 1 → Phase 2 → Phase 3 → Phase 4. No circular dependency. Each phase is a prerequisite for the next and produces a usable artifact.

Key Principles

• Every tool is written in CL, for CL, by the agent
• The 1994 standard is preserved — modernization lives alongside it
• The agent is the primary consumer (human UX is secondary)
• The HOL prover grows from use — every proved theorem is future automation