Convergent Weierstrass division #
The Weierstrass division theorem for convergent power series [GR65, Chapter II, §B],
[Nar66, Chapter II, Theorem 2]: for convergent f, g with g x_0-regular of order d, there
are unique convergent q, r with f = q g + r and r of x_0-degree < d. Existence follows
the Banach-algebra argument [GR71, Kapitel I]: normalize g = E (x_0^d - h) (Normalize.lean),
pick a radius vector where f and h converge, shrink its tail so that ‖h‖_ρ ≤ ρ_0^d / 2
(Splitting.lean), divide by the contraction (Contraction.lean), f = q₁ (x_0^d - h) + r, and
put q := q₁ E⁻¹. Uniqueness: a convergent division datum is a formal Weierstrass division over
the tail ring (isWeierstrassDivision_of_eq in Poly.lean), and Mathlib's formal uniqueness
(PowerSeries.IsWeierstrassDivision.elim) applies. The preparation theorem of
Preparation.lean is the division of x_0^d by g.
The coefficient field K is ℝ or ℂ ([RCLike K]).
Uniqueness of Weierstrass division data, from the formal theorem.
Convergent Weierstrass division: unique convergent q, r with f = q g + r and r of
x_0-degree < d [GR65, Chapter II, §B], [Nar66, Chapter II, Theorem 2].