Result: 4 distinct shapes (saturated rng=2/3 with permute_dedup=False).
The same enumeration was found independently by Scott Vorthmann and David Richter and posted at https://vorth.github.io/vzome-sharing/2007/04/24/five-cell-family.html (2007). Their 4 shapes correspond exactly to ours:
| Their description | Our file |
|---|---|
| Vertex-first (tet + center, all yellow/green) | 5cell_vertex_first_tet_plus_center.vZome |
| Yellow + red + blue (the interactive model) | 5cell_5ball_Y4B2R4.vZome |
| Red + yellow + blue (David’s second) | 5cell_5ball_R6Y1B3.vZome |
| Yellow + blue only (David’s third, 4-ball) | 5cell_4ball_Y6B3.vZome |
Standard ℤ[φ] embedding aligned with the 4th coordinate axis:
(Note √5 = 2φ − 1 ∈ ℤ[φ], so the embedding is in ℤ[φ]⁴.) Edge length² = 200 between every pair.
The 5-cell embedding singles out the 4th axis (no S₄ permutation
symmetry on coordinates), so we run the search with
permute_dedup=False in gen_dirs. With permutation dedup ON, the
search collapses geometrically distinct directions and reports only 3
shapes.
Verified by fitting a 3×4 linear map M to each emitted vZome model and checking M·Mᵀ = s²·I (uniform scale, no stretching). All 4 files are genuine orthographic projections.
| # | balls | strut sig | kernel n (proportional) | description |
|---|---|---|---|---|
| 1 | 5 | Y:4 G:6 | (1−2φ, 1−2φ, 1−2φ, 1) = (−√5, −√5, −√5, 1) | vertex-first: v₁ collapses to centroid of v₂…v₅ → tet + center |
| 2 | 5 | Y:4 R:4 B:2 | (−1, 0, 0, 1) | mixed Y/R/B |
| 3 | 5 | B:3 R:6 Y:1 | (−1, −1+2φ, 1, 1) = (−1, √5, 1, 1) | mixed B/R/Y |
| 4 | 4 | Y:6 B:3 collapsed:1 | (−1, −1, −1, −1+2φ) = (−1, −1, −1, √5) | degenerate (one edge collapses) |
Note: kernels along any of the 5 vertex directions give the “vertex-first tet + center” shape (only one entry, since group-equivalent kernels produce congruent 3D shapes).
➡️ Open this page on GitHub Pages to interact with the 3D models below (the embeds only render when this file is served via GitHub Pages, not in github.com’s markdown preview).
| Shape | File | Method |
|---|---|---|
| #1 | 5cell_vertex_first_tet_plus_center.vZome |
hand-built |
| #2 | (TBD) | — |
| #3 | (TBD) | — |
| #4 | (TBD) | — |
The vertex-first tet+center is trivial to hand-build (integer coords).
The other three shapes require symbolic projection through their kernel
direction; the standard <Polytope4d group="A4"> approach does not
work because vZome’s A4 embedding differs from ours, so kernel
directions in our embedding do not translate to vZome quaternions
without an explicit basis change. Hand-building the remaining three
needs:
try_align rotation symbolically (rotation matrix in ℤ[φ]).This is parked for later.
cd regular_4polytopes\lib
python -c "from search_engine import gen_dirs, search, group_by_shape; \
from polytopes import POLYTOPES; \
dirs = gen_dirs(rng=2, permute_dedup=False); \
V,E = POLYTOPES['5cell'][0][1](); \
hits = search('5cell', V, E, dirs); \
print(len(group_by_shape(hits, V, E)), 'shapes')"
# Output: 4 shapes