feat(ring_theory/witt_vector): witt_sub, a demonstration of is_poly techniques (#5165)

Co-Authored-By: Rob Y. Lewis <rob.y.lewis@gmail.com>

Author: jcommelin

src/ring_theory/witt_vector/is_poly.lean

@@ -617,4 +617,37 @@ attribute [ghost_simps]
       nat.succ_ne_zero nat.add_sub_cancel nat.succ_eq_add_one
       if_true eq_self_iff_true if_false forall_true_iff forall_2_true_iff forall_3_true_iff
 
+/-!
+### Subtraction of Witt vectors
+
+In Lean, subtraction in a ring is by definition equal to `x + -y`.
+For Witt vectors, this means that subtraction is not defined in terms of
+the polynomials `witt_sub p`.
+
+As a demonstration of some of the techniques developed in this file,
+we show by a computation that evaluating `witt_sub p` on the coefficients of `x` and `y`
+gives the coefficients of `x - y`.
+
+For a more powerful demonstration, see `ring_theory/witt_vector/identities.lean`.
+-/
+
+lemma sub_eq (x y : 𝕎 R) :
+  x - y = eval (witt_sub p) ![x, y] :=
+begin
+  apply is_poly₂.ext ((add_is_poly₂).comp_right
+    (neg_is_poly)) ⟨witt_sub p, by intros; refl⟩ _ _ x y,
+  unfreezingI { clear_dependent R }, introsI R _Rcr x y n,
+  simp only [←sub_eq_add_neg, ring_hom.map_sub],
+  symmetry,
+  have := witt_structure_int_prop p (X 0 - X 1 : mv_polynomial (fin 2) ℤ) n,
+  apply_fun (aeval (function.uncurry ![x.coeff, y.coeff])) at this,
+  simp only [aeval_bind₁, alg_hom.map_sub, bind₁_X_right] at this,
+  simp only [aeval_eq_eval₂_hom, eval₂_hom_rename] at this,
+  exact this,
+end
+
+lemma sub_coeff (x y : 𝕎 R) (n : ℕ) :
+  (x - y).coeff n = peval (witt_sub p n) ![x.coeff, y.coeff] :=
+by { rw [sub_eq], refl }
+
 end witt_vector