I brute forced a geometric figure together for a nerdy game thing without the math chops to now what it is geometrically. I'd like to know more about it, or if I've made some error in drawing it out. It's eight Heawood map cells where each face is in two cells and no pair of faces appear in more than one cell. I think each edge touches 3 faces and each vertex touches 6.What is this figure? Is it a coiled bundle of tori? How many dimensions does it exist in most naturally if each torus lies flat?
And here are some maps drawn on the figure, so long as I already put the effort in for em.
>>17100164>>17100167These are awesome maps and you game sounds cool. Each game space existing in two worlds at once is a cool move.I have no idea, so I asked my smart SI fren:Anon has not drawn a bundle of tori. They have drawn, cell‑by‑cell, the cusp‑torus diagram (equivalently the dual special spine) of the Thurston congruence link complement — the 8‑component link whose exterior is tiled by 28 regular ideal hyperbolic tetrahedra, the regular tessellation U2+ζ{3,3,6}\mathcal{U}^{\{3,3,6\}}_{2+\zeta}U2+ζ{3,3,6} with cusp modulus 2+ζ2+\zeta2+ζ and symmetry group PGL(2,7)\mathrm{PGL}(2,7)PGL(2,7) (order 336). Its 2‑skeleton is an immersed punctured Klein quartic. The incidence counts Anon guessed are correct; the only conceptual error is calling it a surface/bundle. And "if each torus lies flat," the natural home is 3 dimensions — but hyperbolic, not Euclidean.https://arxiv.org/html/1406.2827v3https://unhyperbolic.org/thesis.pdfThat means jack shit to me. Want more?
>>17100289#Error checklist for the drawing#Anon's counts are right; verify these to certify the gluing:•Each of the 28 letters appears exactly 12 times (3 corners × 4 cells).•Each pair of cells shares exactly 6 letters (the shared hexagon's corners).•Within each cell, every pair of hexagons shares exactly one edge (2 letters); the 14 vertex‑triples form two disjoint Fano planes.•Each triple of cells occurs as one edge in all three cells with the same two endpoint letters, and is contained in exactly two letters' 4‑sets.•Chirality/orientation: T2+ζT_{2+\zeta}T2+ζ is chiral. The two ends of an edge must read the shared hexagon's neighbor cycle in opposite cyclic orders; if some cells are mirrored inconsistently you will get the mirror manifold or a non‑orientable/invalid spine rather than the Thurston link.If all five pass, congratulations to Anon: a brute‑forced "nerdy game thing" has reconstructed, from the round‑robin K8K_8K8 pattern and seven‑color torus cells, the cusp diagram of one of the most symmetric objects in low‑dimensional topology — the Thurston congruence link, with its Fano planes, equianharmonic design, and immersed Klein quartic. It can be confirmed in SnapPy/Regina by building the 28‑tetrahedron ideal triangulation from the letter data and checking that all edge degrees are 6 and all eight vertex links are the 7‑vertex torus.
>>17100299#The dictionary: your picture an ideal triangulation#Each of your cells is the Heawood map: 7 mutually adjacent hexagons on a torus, 14 vertices, 21 edges — the dual of the minimal 7‑vertex (Möbius–Császár, K7K_7K7) torus triangulation. That is exactly the cusp cross‑section T2+ζT_{2+\zeta}T2+ζ of the Thurston link (14 equilateral triangles, dual 7 hexagons).Reading your figure as the dual spine of an ideal triangulation so "each face in two cells, no pair of faces in more than one cell" is precisely the K8K_8K8 condition, and "each edge touches 3 faces / each vertex touches 6" is the tetrahedron bookkeeping. Every edge has degree 6 (hexagon), which is the defining condition of a tetrahedral manifold: regular ideal tetrahedra, 6×60∘=360∘6\times60^\circ=360^\circ6×60∘=360∘ around every edge. Euler characteristic of your complex: 28−56+28=028-56+28=028−56+28=0, matching a compact 3‑manifold with 8 torus boundary components. Goerner's classification shows the regular tessellation with cusp modulus 2+ζ2+\zeta2+ζ is unique (28 solids, 8 cusps; his category C2+ζ\mathcal{C}_{2+\zeta}C2+ζ has one object) arxiv.org — so if your labels glue consistently, you have drawn that manifold, not a lookalike. Volume =28×1.0149416≈28.418=28\times1.0149416\approx28.418=28×1.0149416≈28.418.The 28 vertex labels are the 28 "equianharmonic" 4‑subsets of P1(F7)\mathbb{P}^1(\mathbb{F}_7)P1(F7) (the PGL(2,7)\mathrm{PGL}(2,7)PGL(2,7)‑orbit with A4A_4A4 stabilizer): a 3-(8,4,2)3\text{-}(8,4,2)3-(8,4,2) design — each letter lies in 4 cells, each triple of cells in exactly 2 letters, and inside each cell the 14 vertex‑triples are two disjoint Fano planes (the two chiralities of corners of the K7K_7K7 torus). Flag count 28×12=336=∣PGL(2,7)∣28\times12=336=|\mathrm{PGL}(2,7)|28×12=336=∣PGL(2,7)∣: the tessellation is flag‑transitive, i.e. regular.
>>17100289I only recognize enough terms to be dangerous. I'm guessing the "more" would be lost on me, but post whatever if it's not a hassle or anything.After all the work, it's not even all that useful for what I was going for. You only have to move at most twice to get from any one place to another, which should have been obvious to me because of how I started with all of the hexes touching and everything. This somehow started as a D&D project. It was going to be a cosmology for a setting or a new version of Planescape or Spelljammer. In the end its a set of pretty pictures, which is alright.
>>17100307Woops. I just hadn't refreshed
>>17100303#What it is, and what it is not#•It is a 2‑dimensional cell complex, not a surface. Three hexagons meeting along every edge means branch lines; the vertex links are K4K_4K4 (4 edge‑germs, 6 face‑corners). This is exactly a Matveev special polyhedron / special spine, locally identical to a soap foam obeying Plateau's laws (3 films per edge, 4 edges per tetrahedral vertex). It is a spine of the compact core of the link complement.•It is not a "coiled bundle of tori." In the actual 3‑manifold the 8 tori are disjoint boundary/cusp tori; they generate H2≅Z7H_2\cong\mathbb{Z}^7H2≅Z7 with the single relation that their sum bounds. The shared hexagon in your diagram does not mean the tori intersect in space — it records the triangulation edge joining those two cusps. As an abstract complex it is a branched plumbing of 8 tori along the K8K_8K8 pattern, which is a perfectly good object, just not a fiber bundle or a manifold.•The "maps" in your second image are legitimate: curves and regions live on this 2‑complex. Note the sum of all 28 hexagons is a Z/3\mathbb{Z}/3Z/3‑cycle (each edge appears 3 times) but not a Z/2\mathbb{Z}/2Z/2‑cycle — another signature that this is a spine, not a closed surface.
>>17100312#How many dimensions, if each torus lies flat?#Three answers, depending on what "flat" and "lies" mean:Natural home: 3‑dimensional hyperbolic space. Each Heawood cell is intrinsically a Euclidean (flat) torus, and flat tori sit perfectly "flat" in H3\mathbb{H}^3H3 as horotori — cusp sections. That is literally what your figure is: the 8 flat cusp tori of a cusped hyperbolic 3‑manifold. So the honest answer is 3, with hyperbolic geometry doing the work that Euclidean geometry cannot.One flat torus in Euclidean space: 4. A smooth metrically flat torus embeds isometrically in R4\mathbb{R}^4R4 (Clifford torus) and not in R3\mathbb{R}^3R3 en.wikipedia.org.Your hexagons, flat and planar, glued as drawn: impossible in every Rn\mathbb{R}^nRn. At each cell vertex three regular hexagon corners give three rays at mutual 120∘120^\circ120∘; three such rays are coplanar in any dimension (their Gram matrix has rank 2), so the three hexagons are forced coplanar, and this propagates across the whole torus — a flat planar torus is a plane. Equivalently: a degree‑3 vertex with cone angle 360∘360^\circ360∘ has a spherical‑triangle link of perimeter 2π2\pi2π, which must be a great circle. This is the rigidity behind Schwartz's recent "paper tori" results (no 7‑vertex flat polyhedral torus in R3\mathbb{R}^3R3; 8 exists) arxiv.org; your 14‑vertex, all‑degree‑3 hexagonal cell is even more obstructed. The glued complex is worse: each vertex has cone angle 6×120∘=720∘6\times120^\circ=720^\circ6×120∘=720∘ and four edges at mutual 120∘120^\circ120∘, which no Euclidean space admits. Topologically the spine does embed in R3\mathbb{R}^3R3 (the link complement retracts onto it), just not with flat regular hexagons. Your quotient/fundamental‑domain drawing is therefore the correct representation — not a failed embedding.
>>17100307lol. No worries. Free token burn. My Chinese AI has gifted me a free-trial for new model to math up.I am high af and you maps are fucking awesome and it’s hysterical not to notice the two-move rule AFTER you put in all the work, but you have incredible ideas.Game on. OPNow I need to talk to you mom alone for a while.
>>17100325Said what i need to your mom, OP.I still have no clue what any of it means but I hope you have fun with it.https://chat.qwen.ai/s/deploy/t_b70363b5-bee4-4ebc-8b31-9b22961b19c4