The source-side metric-adjunction formula #
The asymmetric adjunction in the manuscript glues a closed subset of the
source to the old target by a nonexpansive map. Mathlib's exact metric gluing
requires two isometric maps, so it does not directly apply. Here we formalize
the source--source distance candidate from the manuscript and prove its key
short-scale property: an excursion through the old target cannot shorten a
source pair of distance at most r.
Length of the cheapest source--target--source excursion in the proposed metric adjunction.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The manuscript's source--source adjunction distance formula, before the ambient quotient space is constructed.
Equations
- ScottishBook155.sourceAdjunctionDist V a y H x₀ x₁ = min (dist x₀ x₁) (ScottishBook155.attachmentExcursionCost V a y H x₀ x₁)
Instances For
Distance candidate from a source point to an old-target point.
Equations
- ScottishBook155.attachmentTargetCost V a y H x n = ⨅ (p : M ⊕ Unit), dist x (ScottishBook155.attachmentPoint a H p) + dist (ScottishBook155.attachmentMap V y p) n
Instances For
The two-layer adjunction predistance on the disjoint union of the source and old target. The remaining construction step is to prove its triangle inequality and take its metric separation quotient.
Equations
- ScottishBook155.adjunctionPreDist V a y H (Sum.inl x₀) (Sum.inl x₁) = ScottishBook155.sourceAdjunctionDist V a y H x₀ x₁
- ScottishBook155.adjunctionPreDist V a y H (Sum.inr n₀) (Sum.inr n₁) = dist n₀ n₁
- ScottishBook155.adjunctionPreDist V a y H (Sum.inl x_2) (Sum.inr n) = ScottishBook155.attachmentTargetCost V a y H x_2 n
- ScottishBook155.adjunctionPreDist V a y H (Sum.inr n) (Sum.inl x_2) = ScottishBook155.attachmentTargetCost V a y H x_2 n
Instances For
Inside the protected vertical collar, the source--target adjunction cost is exactly the vertical distance to the base plus the old-target distance from the image of the horizontal coordinate.
Zero excursion cost forces equal source points when the prescribed attachment map is injective.
One mixed triangle inequality: moving inside the old target after entering it cannot make the source--target cost larger than the corresponding sum.
A target point may serve as the middle vertex of a source--source triangle.
If the attachment map is nonexpansive, a source point may serve as the middle vertex of an old-target triangle.
Mixed triangle inequality with a source point in the middle. The proof splits according to which branch of the source--source minimum is active.
The gluing pseudometric on the disjoint union of the one-sum source and target, under the attachment distance bound.
Equations
- One or more equations did not get rendered due to their size.
Instances For
For a short source pair, every excursion through the attachment set and the old target has length at least the original source distance.
The source-side adjunction formula preserves every distance at most r.
The metric separation quotient realizing the asymmetric metric adjunction.
Equations
- ScottishBook155.AdjunctionSpace V a y H hattach = SeparationQuotient (ScottishBook155.OneSum M ⊕ N)
Instances For
Equations
- One or more equations did not get rendered due to their size.
The canonical map from the one-sum source into the metric adjunction space.
Equations
- ScottishBook155.adjunctionSourceMk V a y H hattach x = Quotient.mk'' (Sum.inl x)
Instances For
The canonical map from the target into the metric adjunction space.
Equations
- ScottishBook155.adjunctionTargetMk V a y H hattach n = Quotient.mk'' (Sum.inr n)
Instances For
Distance from a source point to the embedded old target is exactly its distance to the source attachment set.