Talk:Spherical polyhedron

Table of images
I dropped in a table of spherical tiling images I had on Wythoff symbol. Tom Ruen 21:21, 31 March 2007 (UTC)
 * Thanks, Tom. Steelpillow 09:46, 1 April 2007 (UTC)

Rigorous Definition
Nice attractive article, lovely pictures, good intuitive approach. But us theoreticians need a more formal definition also. Also, the article should give non-examples - such as a toroidal polyhedron. I'm not expert enough on this subject (yet). SteveWoolf (talk) 13:06, 14 November 2008 (UTC)


 * Well, other than describing it as a tiling of a sphere I am not sure what else one can say. The terms "tiling" and "sphere" are rigorous enough (unless one invokes higher-dimensional spheres, which is not the case here). Certainly neither Poinsot nor Coxeter gave a less intuitive definition, and I am not aware of any other (that's not to say that someone reputable might have done somewhere). I do not think that non-examples would do anything other than confuse the casual reader, unless you have some rather subtle point to make. -- Cheers, Steelpillow (Talk) 18:00, 14 November 2008 (UTC)


 * I agree this is lacking in details. I think it's valuable (somewhere) to recognize the 3 classes of constant curvature spaces - spherical (K=1), Euclidean (K=0), and hyperbolic (K=-1), with geometric properties that a face can be translated without distortion, while scaling can preserve lengths, but not area. Something like that.
 * A wlink Spherical tiling goes to a tessellation section, which maybe shows (or ought to show) a connection to the wider potential of tiling other surfaces, but underdeveloped there. There's the general idea of tessellations and tilings, and then there's the more specific form topologically related to polyhedra with edge-to-edge connection of faces. The MAIN value I know for spherical polyhedra is the relation to symmetry, and degenerate forms like hosohedrons and dihedrons which exist as tilings on a sphere.
 * It should reference spherical geometry, also underdeveloped article. Perhaps it ought to be clarified here that edges are "straight", following great circle arcs. Also spherical trigonometry is important, if you want to compute information about polygonal faces on a spherical tiling.
 * Another relation I know is the Conway polyhedron notation, which creates topological operators to relate polyhedra, and it was implemented by George Hart for spherical tilings specifically.
 * Tom Ruen (talk) 18:43, 14 November 2008 (UTC)
 * The topics of curved spaces, isometric transformations and tilings generally, all have their own pages: all this page needs for these is a few links, either in the main text or a See also section. However an important aspect of spherical polyhedra does arise from all this - the spherical kaleidoscopes, from which we derive tilings (i.e. spherical polyhedra) of M&ouml;bius triangles and the more general Schwarz triangles, which in turn lead to the convex and star uniform polyhedra respectively. I guess this is probably the symmetry aspect that Tom mentions. -- Cheers, Steelpillow (Talk) 21:51, 14 November 2008 (UTC)
 * Hi again...my point is that as an abstract polytopist, I'd like to see a formal definition - i.e. what is a "sphere" in an abstract context? But I think this highlights a much wider issue here - namely that different polytope enthusiasts have very different aims and perspectives. Many, many are quite content to explore and enumerate all the pols that have some property X - such as Johnson solids. Others enjoy the graphics aspect of it all. These worthy people do not sully themselves with the finer points of theory. And to be sure, their work is useful and extremely attractive, respectively. (And a lot more attractive than 99% of what allegedly savvy art critics consider modern art). Some people like Euclidean spaces - me, I like abstract pols and a fully rigorous treatment. The problem is, abstract pols and their Euclidean relatives are precisely that - related, but not the same.
 * Which means: Some articles need to be SPLIT, with a new one for the abstract case. Others simply need an abstract section in the same  article. For example, nearly all of the Polygon article has little to do with an abstract polygon. On the other hand, we hardly need 2 separate articles for Icosidodecahedron.  This will be a great deal of work, given the never-ending proliferation of polytopic articles. (Actually it's really great - that so many have contributed so much).  But if the articles were worth creating, as indeed they are, then our goal should be to make them RIGHT - and abstract pols, in my view, deserve a lot more recognition than they now have. See ya SteveWoolf (talk) 01:14, 15 November 2008 (UTC)
 * Hi Steve. Your criticisms of "realists" are well founded, and extend even to those who are concerned with the theory - Gr&uuml;nbaum has called this the "original sin" in the theory of polyhedra, and it stretches all the way back from Coxeter to Euclid. This is in fact a major (if not the) problem that abstract poltyopes were developed to address. However in the context of "spherical polyhedra", this term is usually used to distinguish a tiling of a real geometric 2-sphere from its flat-faced cousin; this is a purely (geo)metric distinction and has no abstract significance. So for the main body of this article the introduction of abstract polytopes is not appropriate. There may be an exception, in that abstractly, an orientable polyhedron of Euler characteristic 2 is topologically equivalent to a 2-sphere, and here the term "spherical polytope" might be appropriate. If it has appeared in this context in a reputable source then we can talk about this context in the article, but if it has not then an article titled "spherical polyhedron" is the wrong place to discuss it - the Abstract polytope article would be a better place. -- Cheers, Steelpillow (Talk) 09:41, 15 November 2008 (UTC)
 * I don't think you are right about this, but I'm not sure, this is an area I need to understand better. What seems clear to me is that the cube or square pyramid, e.g., are spherical. Now take 4 thin trapezoid "cubes", and glue them together on their bevilled edges to form a closed circuit with a hole. This is a toroidal polyhedron.
 * Note that the graph of this toroid (i.e. vertexes and edges only) is the same as for a 4-cube. To make it a 4-cube you need two more cells (3-faces) - the inner hole and the "outer" cube; then you remove the old maximal 3-face and replace it with a new maximal 4-face four more cells.SteveWoolf (talk) 22:08, 10 January 2009 (UTC)
 * Anyway, the point is that the "spherical" concept, in my view, is of central importance to abstract theory in classifying polytopes. I'll invite Mike and CG to comment, as they are likely more knowledegable about this. Meanwhile I'll try to improve my understanding. Either way, none of this topic is "new" ground, I am sure. Nice weekend SteveWoolf (talk) 11:46, 15 November 2008 (UTC)
 * I can't say I've worked with this myself, but here is what McMullen and Schulte's book has to say: "An abstract n-polytope P is called spherical if it is isomorphic to the face-lattice of a convex n-polytope Q." (p 153). Similarly, an abstract polytope is locally spherical if each proper section (equivalently all facets and vertex figures) are spherical. They say very little else about general spherical polytopes, though there is somewhat more about regular spherical polytopes. --CunningGabe (talk) 01:16, 17 November 2008 (UTC)

Thanks for that - at least you have confirmed what I intuitively feel to be spherical. However, I feel that we (a) Need to define this is a purely abstract manner, without recourse to classical ancestors, and then, (b) Prove that the abstract version is equivalent to the "cross-cultural" one. I don't think this will be especially difficult, and the exercise itself may produce new insights. I shall think on this.... SteveWoolf (talk) 07:16, 17 November 2008 (UTC)


 * Thanks for the example usage in an abstract context. One thing is certain, if by "cross-cultural" is meant the established usage when discussing spherical geometry, kaleidoscopes, etc. then it is quite distinct from the abstract usage - the one is geometric, the other abstract. It is a not even a necessary condition for any geometrically spherical polyhedron to be topologically (abstractly) spherical, as may be seen for example in the original exhaustive study of Uniform polyhedra:
 * H.S.M. Coxeter, M.S. Longuet-Higgins, J.C.P. Miller, Uniform polyhedra, Phil. Trans. 1954, 246 'A, pp 401-450.
 * where several topologically non-spherical polyhedra are derived from kaleidoscopic tilings of the sphere. However the idea of convexity is a slightly better match - for a polyhedron to be geometrically convex it must be a topological sphere (though the converse is not true).
 * Any remarks on spherical topology are probably best included in a general discussion of topological types. For some starting points, see for example topology and genus. -- Cheers, Steelpillow (Talk) 20:48, 17 November 2008 (UTC)

Okay, I now have a quite elegant abstract defn of spherical! Correct? We'll see. But I think it's better to centralise our abstract talk in one place, so you can see this in Talk:Abstract polytope. SteveWoolf (talk) 04:26, 18 November 2008 (UTC)

Projective polyhedra
I have not seen the term "projective polyhedron" used in this context. Either this needs referencing (e.g. did Coxeter or Gr&uuml;nbaum use it?), or replacing with a more general description. -- Cheers, Steelpillow (Talk) 09:12, 14 April 2010 (UTC)


 * Hi Guy,
 * You raised this and I addressed this on my talk page, but for reference:
 * I’ve referenced with McMullen and Schulte, and the terminology is further discussed at projective polyhedron and Talk:Projective_polyhedron if anyone would like to read or discuss further on this.
 * —Nils von Barth (nbarth) (talk) 00:34, 23 April 2010 (UTC)

Limiting cases
I moved this deleted section here below, topologically valid tessellations, in case someone else can someday find some references that can support them. My only source is Coxeter:
 * Tom Ruen (talk) 20:46, 6 February 2015 (UTC)

-