Skip to content

Repository files navigation

Ia

Ia (pronounced “eye-uh”) is a human- and AI-readable, token-efficient language for encoding procedures executed by people, groups, and AI agents.

Design goals

  • Readable: close to natural procedural language.
  • Minimal: few orthogonal forms and little boilerplate.
  • Agent-explicit: only named Agents execute Functions.
  • Composable: complex procedures decompose into ordered Function Lists.

Status

Ia v15.1.0 and Standard Library v1.0 are the current design specifications. This repository includes a reference parser, a minimal language server, and a VS Code extension. Control flow and execution are intentionally deferred.

Examples and analysis

Language tooling

The first tooling milestone provides:

  • syntax highlighting and comment behavior for .ia files;
  • an error-recovering parser generated from the Ia grammar; and
  • diagnostics for malformed syntax, invalid imports, and duplicate bindings.

Unknown domain Functions remain valid. The tooling does not infer output counts or attempt to execute Ia.

To build and test it:

npm ci --ignore-scripts
npm test

To try the extension, open this repository in VS Code and press F5. The new Extension Development Host recognizes .ia files automatically.

The implementation is split into three small parts: the parser and analyzer in src/language, the language server in src/server.ts, and the VS Code client in src/extension.ts.

Research

Next work

  1. Design control flow as Agent-executed built-in Functions.
  2. Add symbol navigation and import-aware completion after the parser surface stabilizes.
  3. Define executor profiles and conformance tests.

About

A human- and AI-readable, token-efficient pseudocode language for encoding procedures.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages