Lean migration: Phase 4 (IterProd + FiniteMap lattices)
- Spa.Lattice.IterProd: k-fold product, recursive Lattice instance,
fixed height k*hA + hB, bot = build of bottoms
- Spa.Lattice.FiniteMap: spine-pinned assoc lists ({l // l.map fst = ks});
with = the 1100-line Map.agda collapses into positional 'combine'.
Same lemma inventory (membership, locate, updating, GeneralizedUpdate,
valuesAt, Provenance-union, le_of_mem_mem) — Nodup is now an explicit
hypothesis where the Agda Map carried it intrinsically. Fixed height
|ks|*hB still via transport along the IterProd isomorphism, which no
longer needs Unique ks (representation is canonical).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-06-09 19:12:39 -07:00
|
|
|
|
/-
|
|
|
|
|
|
Port of `Lattice/IterProd.agda`: the `k`-fold product `A × (A × ⋯ × B)`.
|
|
|
|
|
|
|
|
|
|
|
|
With propositional equality and typeclasses, the Agda `Everything` record
|
|
|
|
|
|
(which threaded the lattice operations and the conditional fixed-height proof
|
|
|
|
|
|
through one recursion, so that the operations built by separate recursions
|
|
|
|
|
|
would agree) is no longer needed: the `Lattice` instance is one recursive
|
|
|
|
|
|
definition, and the fixed-height structure is another recursion over it.
|
|
|
|
|
|
|
|
|
|
|
|
Correspondence:
|
|
|
|
|
|
IterProd ↦ Spa.IterProd
|
|
|
|
|
|
build ↦ Spa.IterProd.build
|
|
|
|
|
|
isLattice/lattice ↦ instance Spa.IterProd.instLattice
|
|
|
|
|
|
fixedHeight,
|
|
|
|
|
|
isFiniteHeightLattice,
|
2026-06-22 18:33:48 -05:00
|
|
|
|
finiteHeightLattice ↦ Spa.IterProd.fixedHeight (+ instFiniteHeight instance)
|
Lean migration: Phase 4 (IterProd + FiniteMap lattices)
- Spa.Lattice.IterProd: k-fold product, recursive Lattice instance,
fixed height k*hA + hB, bot = build of bottoms
- Spa.Lattice.FiniteMap: spine-pinned assoc lists ({l // l.map fst = ks});
with = the 1100-line Map.agda collapses into positional 'combine'.
Same lemma inventory (membership, locate, updating, GeneralizedUpdate,
valuesAt, Provenance-union, le_of_mem_mem) — Nodup is now an explicit
hypothesis where the Agda Map carried it intrinsically. Fixed height
|ks|*hB still via transport along the IterProd isomorphism, which no
longer needs Unique ks (representation is canonical).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-06-09 19:12:39 -07:00
|
|
|
|
⊥-built ↦ Spa.IterProd.bot_fixedHeight
|
|
|
|
|
|
-/
|
|
|
|
|
|
import Spa.Lattice.Prod
|
|
|
|
|
|
import Spa.Lattice.Unit
|
|
|
|
|
|
|
|
|
|
|
|
namespace Spa
|
|
|
|
|
|
|
|
|
|
|
|
universe u
|
|
|
|
|
|
|
|
|
|
|
|
/-- Agda: `IterProd k = iterate k (A × ·) B`. (As in the Agda module, `A` and
|
|
|
|
|
|
`B` are constrained to the same universe to keep the recursion simple.) -/
|
|
|
|
|
|
def IterProd (A B : Type u) : ℕ → Type u
|
|
|
|
|
|
| 0 => B
|
|
|
|
|
|
| k + 1 => A × IterProd A B k
|
|
|
|
|
|
|
|
|
|
|
|
namespace IterProd
|
|
|
|
|
|
|
|
|
|
|
|
variable {A B : Type u}
|
|
|
|
|
|
|
|
|
|
|
|
instance instLattice [Lattice A] [Lattice B] :
|
|
|
|
|
|
∀ k, Lattice (IterProd A B k)
|
|
|
|
|
|
| 0 => inferInstanceAs (Lattice B)
|
|
|
|
|
|
| k + 1 => @Prod.instLattice A (IterProd A B k) _ (instLattice k)
|
|
|
|
|
|
|
|
|
|
|
|
instance instDecidableEq [DecidableEq A] [DecidableEq B] :
|
|
|
|
|
|
∀ k, DecidableEq (IterProd A B k)
|
|
|
|
|
|
| 0 => inferInstanceAs (DecidableEq B)
|
|
|
|
|
|
| k + 1 => @instDecidableEqProd A (IterProd A B k) _ (instDecidableEq k)
|
|
|
|
|
|
|
|
|
|
|
|
/-- Agda: `build`. -/
|
|
|
|
|
|
def build (a : A) (b : B) : (k : ℕ) → IterProd A B k
|
|
|
|
|
|
| 0 => b
|
|
|
|
|
|
| k + 1 => (a, build a b k)
|
|
|
|
|
|
|
|
|
|
|
|
variable [Lattice A] [Lattice B]
|
|
|
|
|
|
|
2026-06-22 18:33:48 -05:00
|
|
|
|
def fixedHeight [FiniteHeightLattice A] [FiniteHeightLattice B] :
|
|
|
|
|
|
∀ k, FiniteHeightLattice (IterProd A B k)
|
|
|
|
|
|
| 0 => inferInstanceAs (FiniteHeightLattice B)
|
|
|
|
|
|
| k + 1 => @Spa.prod A (IterProd A B k) _ (instLattice k) _ (fixedHeight k)
|
Lean migration: Phase 4 (IterProd + FiniteMap lattices)
- Spa.Lattice.IterProd: k-fold product, recursive Lattice instance,
fixed height k*hA + hB, bot = build of bottoms
- Spa.Lattice.FiniteMap: spine-pinned assoc lists ({l // l.map fst = ks});
with = the 1100-line Map.agda collapses into positional 'combine'.
Same lemma inventory (membership, locate, updating, GeneralizedUpdate,
valuesAt, Provenance-union, le_of_mem_mem) — Nodup is now an explicit
hypothesis where the Agda Map carried it intrinsically. Fixed height
|ks|*hB still via transport along the IterProd isomorphism, which no
longer needs Unique ks (representation is canonical).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-06-09 19:12:39 -07:00
|
|
|
|
|
2026-06-22 18:33:48 -05:00
|
|
|
|
instance instFiniteHeight [FiniteHeightLattice A] [FiniteHeightLattice B] (k : ℕ) :
|
|
|
|
|
|
FiniteHeightLattice (IterProd A B k) := fixedHeight k
|
|
|
|
|
|
|
|
|
|
|
|
theorem bot_fixedHeight [FiniteHeightLattice A] [FiniteHeightLattice B] :
|
|
|
|
|
|
∀ k, (fixedHeight (A := A) (B := B) k).bot = build (⊥ : A) (⊥ : B) k
|
Lean migration: Phase 4 (IterProd + FiniteMap lattices)
- Spa.Lattice.IterProd: k-fold product, recursive Lattice instance,
fixed height k*hA + hB, bot = build of bottoms
- Spa.Lattice.FiniteMap: spine-pinned assoc lists ({l // l.map fst = ks});
with = the 1100-line Map.agda collapses into positional 'combine'.
Same lemma inventory (membership, locate, updating, GeneralizedUpdate,
valuesAt, Provenance-union, le_of_mem_mem) — Nodup is now an explicit
hypothesis where the Agda Map carried it intrinsically. Fixed height
|ks|*hB still via transport along the IterProd isomorphism, which no
longer needs Unique ks (representation is canonical).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-06-09 19:12:39 -07:00
|
|
|
|
| 0 => rfl
|
|
|
|
|
|
| k + 1 => by
|
2026-06-22 18:33:48 -05:00
|
|
|
|
show ((⊥ : A), (fixedHeight (A := A) (B := B) k).bot)
|
|
|
|
|
|
= ((⊥ : A), build (⊥ : A) (⊥ : B) k)
|
|
|
|
|
|
rw [bot_fixedHeight k]
|
Lean migration: Phase 4 (IterProd + FiniteMap lattices)
- Spa.Lattice.IterProd: k-fold product, recursive Lattice instance,
fixed height k*hA + hB, bot = build of bottoms
- Spa.Lattice.FiniteMap: spine-pinned assoc lists ({l // l.map fst = ks});
with = the 1100-line Map.agda collapses into positional 'combine'.
Same lemma inventory (membership, locate, updating, GeneralizedUpdate,
valuesAt, Provenance-union, le_of_mem_mem) — Nodup is now an explicit
hypothesis where the Agda Map carried it intrinsically. Fixed height
|ks|*hB still via transport along the IterProd isomorphism, which no
longer needs Unique ks (representation is canonical).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-06-09 19:12:39 -07:00
|
|
|
|
|
|
|
|
|
|
end IterProd
|
|
|
|
|
|
|
|
|
|
|
|
end Spa
|