Polytope operations

An operation is a kind of transformation that can be applied to a polytope. It can be viewed as a function whose input and output are both polytopes.

This is an incomplete list of operations.

Wythoffian operations
Wythoffian operations encompass all the polytopes obtainable by ringing different nodes of a polytope's Coxeter-Dynkin diagram, as well as some degenerate cases for branched, looped, or non-integer-weighted diagrams. These operations are typically done to regular polytopes to produce Wythoffian uniform polytopes, although they may be done to uniform polytopes and produce degenerates (as seen in some of sidtid's other ringings, for example).

Alternation


Alternation basically removes every other vertex of a polytope with even-sided faces (note that this only holds true for convex finite polytopes). New edges are then added between remaining vertices that shared a removed vertex.

Dual


The dual of a polytope has vertices corresponding to the original polytope's facets, edges corresponding to the original's ridges (in 3D, edges), faces corresponding to the original's peaks (in 3D, vertices), and so on, culminating in facets corresponding to the original's vertices.

The dual of the dual is the original polytope. A polytope may be its own dual.

Petrie dual


In 3D, one can make specific paths along the edges of polyhedra that form skew polygons. Several of these skew polygons together can constitute a polyhedron that looks much like the original, but has skew faces. Doing this is known as the "Petrie dual" or "Petrial" operation.

Like the dual, the petrial of the petrial is the original polyhedron.

This operation may generalize to other dimensions.

Conjugation
Mathematically, replacing a polytope's vertex coordinates with their actual numerical conjugates can produce the polytope's conjugate.

Visually, this usually means switching pentagons with pentagrams, octagons with octagrams, and so on, producing a polytope with the same element counts but that may look radically different. Triangles and squares are unaffected. Polytopes are not stellated into compound forms (such as the hexagram or the {8/2} compound of two squares).

More than two polytopes may be present in a field of conjugates. 4D m-n duoprisms for odd (>3) values of m or n typically have many conjugates, and the uniform polyhedra snid, gosid, gisid, and girsid are all conjugates of one another.

Compounding
A compound of polytopes consists of multiple polytopes made to overlap with one another, often in a symmetrical way.

The compounds themselves are not typically regarded as polytopes, because they are not fully connected.

Blending
Blending can be considered a special case of compounding, where the component polytopes' elements are made to coincide. Duplicate coincident elements are removed to obtain the final result.

Blends can be classified as "inner" or "outer." Inner blends typically involve many copies of a polytope inside the convex hull of a larger polytope, usually creating uniform polytopes with many self-intersections. Outer blends, however, are more similar to the "augmentation" process found in the Johnson solids and other CRF polytopes, and loosely speaking involve gluing two polytopes together at a facet.

Faceting


Faceting a polytope involves finding coplanar combinations of its vertices to create new facets with, not restricted to the outside of the polytope, but able to go through the middle and self-intersect freely. These new facets must then "close up" in a manner befitting a polytope.

Faceting maintains the vertices of the original polytope.

Stellation


Generally, stellation is an operation that extends the facets of a polytope until they once again intersect one another. Its name comes from the star-like results it produces.

Stellation maintains the facet counts of the original polytope, even if each facet is composed of disjoint pieces (as occurs in some of the 59 stellations of the icosahedron).

Some simple stellations may be mistaken for augmentations of pyramids onto the faces of a polytope.

It can help to specify the dimension of the elements being affected by this operation. Thus, the following names were proposed:

Chamfering
Chamfering is similar to the Wythoffian operation of expansion, and like expansion, it may be generalized to higher dimensions. Instead of adding n-cubic facets along the old ridges as is done in expansion, elongated (n-1)-cubic bipyramids are added instead. (In 3D, this translates to adding rectangular-symmetric hexagons at the old edges instead of rectangles or squares.) Facets equivalent to the original vertex figure are not produced (as they are in expansion), but new vertices with a similar vertex figure to the original are.

The result of chamfering also looks similar to adding short, wide frustrums to the faces of the original polyhedron, as long as one ignores the edges that bisect the hexagons this way.

Since chamfering does not produce uniform polytopes, it is not as widely used as similar Wythoffian operations, although it can produce Goldberg polyhedra and even near-miss CRF polytopes.

Goldberg polyhedron


A Goldberg polyhedron is like a repeated chamfering of a polyhedron (usually but not necessarily a Platonic solid), separating the original faces by a great distance and filling in the gaps with many rectangular-symmetric hexagons. They and their duals resemble geodesic domes or insects' compound eyes.

Some Goldberg polyhedra are chiral and cannot be obtained simply by repeated chamfering, though.

Like chamfering, the concept of Goldberg polytopes may be generalized to higher dimensions by using distorted elongated hypercubic bipyramids instead of hexagons.

Zonohedrification
The zonohedrification operation generates a new polytope based on certain vectors, which are in turn based on the position of the vertices of the original polytope relative to its center.

The operation is not widely used, likely because it can be confusing due to its several steps and non-bijectivity (the zonohedrified tetrahedron and zonohedrified cube are both rhombic dodecahedra).

Vertex figure
The act of finding the vertex figure can be thought of as an operation, because it takes in a polytope ("What is the vertex figure of X?") and returns one ("The vertex figure is Y"), although they are not of the same dimension.

Since a non-vertex-transitive polytope will have multiple vertex figures (except in such cases as the pseudorhombicuboctahedron), the "Vertex figure" function would have multiple outputs for a single input, and thus cannot be considered a proper function.

Vertex figures can provide crude visualizations for higher-dimensional polytopes, although this will help few but an expert.