The normal-crossings form of the toric boundary #
The boundary of the analytic realization of a regular fan is the union of the ray-indexed
components TauCeti.Toric.Fan.analyticBoundaryComponent. In the affine chart of a cone σ, read
in the regular coordinates of an integral basis extending the primitive ray generators of σ,
these components are coordinate hyperplanes: a point of the chart lies in the component of a ray
ρ exactly when ρ is a ray of σ and the coordinate indexed by ρ vanishes. Components of rays
outside σ miss the chart.
Taking for σ the cone whose torus orbit contains a given point, and identifying the regular
coordinates linearly with ℂ ^ n, this gives the simple-normal-crossings normal form of the
boundary: near every point there is a holomorphic chart of the realization, defined on the whole
affine chart of σ and biholomorphic onto its image, together with an injection from the
components through the point to the coordinate indices, such that a point of the chart lies in a
component exactly when that component passes through the given point and the corresponding
coordinate vanishes. In particular every boundary component is a closed complex hypersurface,
locally a coordinate hyperplane.
Main declarations #
TauCeti.Toric.Fan.analyticAffineChartι_mem_analyticBoundaryComponent_iff: in the regular coordinates of an affine chart, a boundary component is the vanishing locus of the coordinate of its ray.TauCeti.Toric.Fan.exists_partialDiffeomorph_analyticBoundaryComponent_normalForm: the toric boundary has the holomorphic coordinate-hyperplane normal form near every point.
References #
- W. Fulton, Introduction to Toric Varieties, §§2.1 and 3.1.
- D. Cox, J. Little and H. Schenck, Toric Varieties, §3.2 and §4.1.
In the regular coordinates of the affine chart of a cone σ, a point lies in the boundary
component of a ray ρ exactly when ρ is a ray of σ and the coordinate indexed by ρ
vanishes.
The toric boundary has the holomorphic coordinate-hyperplane normal form near every point x
of the realization. There are a biholomorphism e from an open neighbourhood of x onto an open
subset of ℂ ^ n, for n the rank of the lattice, and an injection j from the set s of
boundary components through x to the coordinate indices, such that a point of the neighbourhood
lies in a boundary component exactly when that component passes through x and the coordinate
it is assigned vanishes.