[meeting] [joint] 26.04.2024

[deemp]

1. Adoption should use B, not n_2? (to fix at least 3 counterexamples)
2. R_COPY seems weird, why do we compute anything inside a formation?
3. How should open recursion be modelled with ρ? (see below)

class A {
  Object f() { return this.g(); }
  Object g() { return 1; }
}

class B extends A {
  Object g() { return 2; }
}

class B1 {
  Object f() { return this.g(); }
  Object g() { return 2; }
}

class Main {
  void main() {
    A x = new B();
    x.f();
  }
}

⟦ t ↦ ⟦
    a ↦ ⟦ b ↦ ξ.ρ.c ⟧,
    c ↦ ∅
  ⟧,
  z ↦ t(c ↦ 42).a.b
⟧.z

⟦ t ↦ ⟦
    a ↦ ⟦ b ↦ ξ.ρ.c ⟧,
    c ↦ ∅
  ⟧,
  z ↦ ⟦ b ↦ ξ.ρ.c, ρ ↦ ⟦ c ↦ 42 ⟧ ⟧.b
⟧.z

⟦
  x ↦ 1,
  a ↦ ⟦ b ↦ ⟦ x ↦ ξ.ρ.x ⟧, x ↦ 2 ⟧,
  c ↦ ξ.a.b.x
⟧

4. Work in progress:

• default atoms (used/generated by EO)
  • Add more built-in dataization functions #291

5. Reminders:

• we currently add λ → Package to program; should this be handled by eoc instead? maybe λ → Test for test files? (yes, create issue for EO)
  • Fix: no top-level lamdba package in generated PHI programs eo#3129 (comment)

6. TODO:

• α0 instead of as-bytes for positional arguments; should this be handled by the normalizer or by eoc phi? (need to implement in normalizer)
  • Support positional arguments #303
• metrics