Session 3 — Loops, main.py & First Running Program¶

What Was Built¶

Loop constructs, the command-line runner, and the first real Terse program executed from a .trs file.

Each Loop¶

Iterates over all nodes of a given type in the knowledge base.

know dog is animal
know cat is animal
know eagle is animal

each creature in animal
  infer creature

The interpreter finds all nodes where is == animal — dog, cat, eagle — and runs the loop body for each, with creature bound to each node in turn.

How it works¶

execute_each() scans facts for all nodes matching the collection type, then runs the body statements once per match with the loop variable added to scope.

Design Decision

each x in animals matches by type — it finds all nodes where is == animals. This mirrors how NCI queries concept clusters — find everything related to a concept and process each one.

While Loop¶

Repeat while a condition holds.

while running is true
  infer model

The interpreter checks facts["running"]["is"] == "true" before each iteration. If true, runs the body. Repeats until false or the 100 iteration safety limit is reached.

Safety Limit

While loops are capped at 100 iterations during Phase 1 development to prevent infinite loops during testing. This limit will be removed or made configurable in a future phase.

Status

While loop parsing and interpreter execution are both implemented. End-to-end testing with a .trs file is planned for Session 5.

main.py — Running .trs Files¶

File: src/interpreter/main.py

The command-line runner. Reads a .trs file and runs it through the full pipeline.

python main.py ../../examples/hello.trs

run_file(filepath) opens the file, passes the source through the Lexer → Parser → Interpreter pipeline, and prints the results.

This is the moment Terse stopped being just code and became a runnable language.

First Real Terse Program — hello.trs¶

File: examples/hello.trs

The first real Terse program executed from a file. A demonstration of facts, relationships, inference rules, functions, and loops all working together.

// Hello World in Terse
// My first real Terse program
// Goatface Tech — 2026

know world is place
know world has life
know life is beautiful

know dog is animal
know dog has fur
know cat is animal
know cat has fur
know bird is animal

when has fur then is mammal
when is place then is real

each creature in animal
  infer creature

each thing in place
  infer thing

Running this file:

python main.py ../../examples/hello.trs

Produces inference results for all animals — dog and cat derive is mammal, bird stays as animal. World derives is real because it is a place.

Phase 1 Progress After Session 3¶

Component	Status
Lexer	✅ Complete
Parser	✅ Complete
Interpreter core	✅ Complete
Facts & relationships	✅ Complete
Inference rules	✅ Complete
Functions with scope	✅ Complete
Markov chain learning	✅ Complete
Each loops	✅ Complete
While loops (parser)	✅ Complete
While loops (e2e test)	🔵 Next session
main.py	✅ Complete
First .trs program	✅ Complete
REPL	✅ Complete
Error handling	⚪ Planned
Tensors	⚪ Planned
Compression keywords	⚪ Planned
Ethics rule syntax	⚪ Planned

Interpreter completeness: ~75%