whetstone_DSL/sprint54_plan.md

Sprint 54 Plan: Homoiconic and AST-Native Family (Lisp, Scheme, Elisp, Smalltalk)

Context

These languages are AST-friendly but carry macro/eval/message semantics that can still break naive projection. Sprint 54 treats them as first-class paradigm targets.

---

Goals

1. Build canonical S-expression/message-send lowering model
2. Preserve macro intent and expansion boundaries in IR
3. Add Smalltalk object/message semantics adapter
4. Validate claim that AST-native authoring projects with low friction

---

Steps

Step 769: S-expression canonical lowering layer (12 tests)

Step 770: Lisp lowering/raising adapters (10 tests)

Step 771: Scheme lowering/raising adapters (10 tests)

Step 772: Elisp lowering/raising adapters (10 tests)

Step 773: Smalltalk lowering/raising adapters (10 tests)

Step 774: Macro boundary + hygienic expansion packet model (10 tests)

Step 775: Eval/runtime boundary risk model (8 tests)

Step 776: whetstone_transpile_ast_native_family MCP tool (8 tests)

Step 777: AST-native projection benchmark suite (8 tests)

Step 778: Sprint 54 integration summary + regression (8 tests)

---

Success Signal

• AST-native authored corpus should achieve highest equivalence confidence and lowest review-required rate among all families.