Genus-two seeds for two-pole gluing #
The two-pole compatibility problem is most naturally studied after retaining
one cross-edge and contracting its separating bridge. The result is a
genus-two/genus-two vertex wedge with glue vertex a. Two uniform facts hold
there without any metric or combinatorial case split:
- the sum
K_A + K_Bof the factor canonical divisors has rank at least one; - the four-chip glue pile has rank at least one and
4a - 2uis winnable for every vertexuof the wedge.
The second statement is the marked seed. Its proof is just the critical
degree form of Riemann's inequality on whichever genus-two factor contains
u. The only remaining issue after restoring the second cross-edge is to
choose one seam phase compatible with the desired rank tests; that scalar
problem is represented by TwoPoleProfile.lean.
A divisor whose degree equals the genus is winnable.
Degree of an integral pile at one vertex.
The local 4a - 2u lemma. On a connected genus-two graph, four
chips at any anchor absorb a doubled chip at any marked vertex.
The four-chip pile on a genus-two/genus-two wedge #
The four-chip pile at the common vertex of a wedge.
Equations
- Utilities.TwoPole.wedgeGluePile G H x y = Utilities.wedgeAddDivisor G H x y (4 • oneChip x) 0
Instances For
The glue pile has the same presentation from the right factor.
The wedge 4a - 2u lemma. Four chips at the glue vertex of a
genus-two/genus-two wedge absorb a doubled chip at every vertex.
Four chips at the glue vertex of a genus-two/genus-two wedge form a rank-one divisor.
The unmarked local-canonical seed #
Unmarked K_A + K_B lemma on the one-pole model. The sum of the two
factor canonical divisors has rank at least one on a wedge of connected
genus-two graphs.