New definition of spherate?

Discussion of shapes with curves and holes in various dimensions.

New definition of spherate?

Postby SharkRetriver » Thu Jan 31, 2013 3:07 pm

Hi, I'm trying to learn about toratopes and got a little, I mean very, confused over what spherate means.
From searching on the forum, it seems to mean "replace every single point in a n-polytope with a n-sphere".
From the toratopic notation page and the list of toratopes, ((II)I) apparently means a torus. This is how I comprehend it.
II makes a square
(II) turns it into a circle, but if every point on a square were replaced by a circle, it would look like a squircle.
(II)I extrudes that circle into a cylinder.
((II)I) somehow transforms that cylinder into a torus? (From the definition given above, a spherated cylinder will be a pillowed cylinder...

I don't get how the definition of spherate matches with what I think is happening in the toratopic notation, and since I'm really a beginner at this (visually learning rather than sticking with mathematical equations) diagrams would be a great help to me.

Thanks in advance!
Posts: 19
Joined: Fri Sep 02, 2011 6:33 pm

Re: New definition of spherate?

Postby wendy » Fri Feb 01, 2013 8:11 am

Spheration is indeed replacing points by spheres, such as might be seen at The points and edges of a body-centred cube have been spherated at different rates.

Whatever the dimensionality of an element, its aroundings (or orthogonal space) is imaged as a sphere, and the element is replaced by an element X sphere prism.
The dream you dream alone is only a dream
the dream we dream together is reality.
User avatar
Posts: 1820
Joined: Tue Jan 18, 2005 12:42 pm
Location: Brisbane, Australia

Re: New definition of spherate?

Postby quickfur » Fri Feb 01, 2013 11:59 pm

I think there's some ambiguity in toratopic notation as to what spheration means. I believe the toratopic notations spheration is based on wendy's rss operator, in which rss(x,y) = sqrt(x^2 + y^2). Manipulating this at the equation level allows you to turn an n-cube symmetry subelement into a circular shape.

As an example of how these operators work, consider the simple case:
|x| + |y| + |z| ≤ 1
If you pick out all the points that satisfies this inequality, you get an octahedron. Now suppose you modify it to be this instead:
max(|x|, |y|, |z|) ≤ 1
Then you get a cube. The idea behind the rss operator, then, is to use rss(x,y,z) instead of max or just sum:
rss(x, y, z) = x^2 + y^2 + z^2 ≤ 1
This is obviously the equation of a sphere. But what if we apply rss to a subset of the variables? Such as:
rss(x,y) = sqrt(x^2 + y^2) ≤ 1
This is a cylinder, if you restrict |z|≤1. But you can do more interesting things too:
sum(rss(x,y),z) = sqrt(x^2 + y^2) + |z| ≤ 1
This produces a double cone (two cones joined base to base, or a rotated diamond, if you will).

IOW, by applying different operations to x, y, z, we can produce a variety of interesting shapes. You can actually experiment with this in povray easily, using the isosurface primitive.

In any case, this is not the same as wendy's definition of spheration.

But I could be misunderstanding it. I've to admit I'm no expert in toratopic notation.
Posts: 2487
Joined: Thu Sep 02, 2004 11:20 pm
Location: The Great White North

Re: New definition of spherate?

Postby wendy » Sat Feb 02, 2013 8:05 am

"Spherate" is a term i invented, to describe what happens when you replace points with balls, and edges with pipes. It includes figures with odd sails representing faces as well. In essence, one replaces thin things with things with a bit of substance for modelling and clarity. The Zome-tool thing makes spherated edge-frames of models, consists of balls which you stick rods into.

Since a circle-edge is a line, you can 'spherate' this too. In the extreme case, you get a thin hoop, but making the spherate larger, the thing becomes more like a car tyre or doughnut. But a torus, be it hoop or doughnut, is itself a solid of interest.

In higher dimensions, even 2-dimensional surfaces are not bounding, so i call these 'hedra' rather than words that suggest division, like 'plane'. You can take a thing like a circle, or a sphere-surface, and spherate it. This is just as in 3d, running a little ball around the figure, and anything the sphere touches is inside it. So a circle perpendicular to a plane in 4d (ie a 3d space), strikes it at two points. If we spherate these two points, we replace these with two little spheres. Rotating this around the whole circle, we get a 'sphere * line' prism, that has been hooked up at the ends, like a hose.

Spheration is a kind of 'modeller's finish', like a paint, applied to thin abstract objects. Other kinds of 'paint' include surtope paint, which replaces smooth surfaces with faces, edges and vertices of a polyhedron. For example, a cube is a sphere painted with surtope paint. More coats of this paint gives more polytope, eg it could come out with hundreds of elements.


The things like rss etc, belong to the domain of the five products. Spheration is not one of them.

The five products produce the five infinite families of regular solids. Three of these five products have expressions in terms of a radiant function, and four are in expressions of adding extra elements to the solid before doing the product.

  • pyramid product, produces simplexes, by draught of content, adds a dimension to the product
  • tegum product (sum), produces the cross polytope, by draught of surface
  • prism product (max), produces the measure polytope, by repetition of content
  • crind product (rss), produces spheres
  • comb product, produces cubic-lattices, by repetition of surface, reduces a dimension from the product

The radiant products work like this. A solid is regarded as a radial function from some centre, where the centre is taken as zero, and the surface is taken as one. This sets a length measure in every direction. When the radiant product is applied over figures X, Y, Z, ..., these figures are placed orthogonally to each other, surface of the product is read as 1 = f(x,y,z, ... ), where f() is the sum, max, or rss functions as the product is the tegum, cube, or sphere.

The standard line goes from -1 to 1 in 1D. The centre of the line is 0, so the radiant function of the line is then abs(x).

If one takes the product of three standard lines in 3d, one gets the standard tegum, cube and sphere respectively. At a point (x,y,z), the radiant coordinate is what (x), (y), and (z) would become, ie abs(x), abs(y), abs(z). The surface of an octahedron is then defined by the surface sum(abs(x), abs(y),abs(z)) = 1. This is the plane x+y+z=1, x-y+z=1, etc. The surface of a cube is defined by max(abs(x), abs(y), abs(z)) = 1, which means that it's bound by planes x=1, x=-1, y=1, y=-1, z=1, z=-1, and the vertices are +/- 1, +/-1, +/- 1. The sphere is defined by rss(abs(x), abs(y), abs(z)) = 1, ie x²+y²+z² = 1².

All products can be freely applied, so one can for example, take the crind-product of x,y and then the prism product of XY, Z. The surface is then described by the general equation 1 = max( rss(abs(x), abs(y)), abs(z)).

These products form a class of 'brick' polytopes, when applied over the standard line. The products are greatly simplified to <tegum>, [prism], (crind), so the standard octahedron is (x,y,z), the standard cube is [x,y,z], and the standard sphere is (x,y,z). A cylinder can be written as [x,(y,z)]. The intersection of two cylinders [(x,z),y] and [(y,z),x], gives a standard crind ([x,y],z).


Torotopes are varieties of round things with holes in them. It's a rather messy process, but involves both the comb product and/or spheration.

The comb product is a repetition of surface. A polygon or circle has a line as a surface. If you multiply a line by a line, you get a rectangle. A polytope bounded by a rectangle is a 3d thing. Let's suppose that you take a sheet of paper. This represents the product of a circle-surface by a circle-edge. You can now hook the left and right hand side of the paper together, to get some sort of hollow cylinder. Now hook the top and bottom, to get a torus.

In four dimensions, the paper has three axies. So after the first two steps, you still have a 'torus-prism'. You can hook the top and bottom together to get a new kind of torus.

The comb product is a product of surfaces, so you are not restricted to surfaces of polygons. You can use polyhedra, etc, as long as the surface dimensions add up to the dimension of the surface of a 4d thing (ie 3). So you can make a dodecahedron - line prism in 4d. We're only interested in the product of surfaces, the pentagonal prisms that make the 'height', not the dodecahedral end-pieces. So we might put a hoop, a thin circle through the dodecahedron faces.

You can then 'bend' the line into a circle, so you get a dodecahedral-circle comb. This would be like stretching it so it goes right around the hoop. This is topological different to the torotopes we met before.

Instead of bending it, you could 'roll it down' as you might take a sock off. The top still meets the bottom, but because it is rolling down outside the prism, the hoop is not covered, and so when the ends hook up, the body of the solid does not include the hoop. It's topologically equal to the previous shape's surface, but the interior is a different shape.

The hose and sock process are identical, but in different directions. So, for example, T( dodeca , line), can be formed by doing a hose process around the dodecahedron, or doing a sock process around the line. The torotope notation is based on this.

Instead of doing a comb product, one can do a series of spherations. For example, the dodecahedron-circle comb, starts off as a point in 4d. Make a sphere around the point, in a 3d space (this spherates the point in a 3-space). Now use a smaller radius, and spherate the surface in 3d + 2d - 1d = 4d. This replaces a point on the surface of a sphere, by a little circle, in the same way that the section of a torus is two small circles.

Going the other way, spherate a point in a 2-space, to get a circle. Now, in 4d, you can spherate the circle, by putting a little sphere at that point.
The dream you dream alone is only a dream
the dream we dream together is reality.
User avatar
Posts: 1820
Joined: Tue Jan 18, 2005 12:42 pm
Location: Brisbane, Australia

Re: New definition of spherate?

Postby Keiji » Sun Feb 03, 2013 3:39 pm

SharkRetriver wrote:Hi, I'm trying to learn about toratopes and got a little, I mean very, confused over what spherate means.
From searching on the forum, it seems to mean "replace every single point in a n-polytope with a n-sphere".

Yes, it does. However, the use of spheration in relation to toratopic notation isn't written up very well.

In order for spheration to make any sense where toratopic notation is concerned, there are two rules of thumb:
1. a parenthesised group of n I's is just an n-sphere, and no spheration actually happens.
2. a parenthesised group containing one (parenthesised) subgroup and n I's not in that subgroup is that subgroup spherated with an (n+1)-sphere.

I don't remember whether spheration actually applies in any sense at all where multiple subgroups are concerned (the tigric aka non-linear toratopes), but I'm not the one who defined spheration and I never found easy to understand myself either.

Rather than concerning oneself with spheration, it is a lot easier to think about it as a parametric equation. In any closed toratope (i.e. wholly contained in an outermost pair of parentheses), the parentheses and I's are really encoding a surface equation with all the squares, square roots and plusses removed - see Rotopic_digit_and_group_notations#To_surface_equation. For open toratopes, these are just Cartesian products of hypercubes and closed toratopes.
User avatar
Posts: 1962
Joined: Mon Nov 10, 2003 6:33 pm
Location: Torquay, England

Re: New definition of spherate?

Postby wendy » Mon Feb 04, 2013 6:44 am

The trick about 'replacing a point with a sphere', will generate intersecting spheres. It just means that the flesh of the bit in question fills in some distance from the bone.
The dream you dream alone is only a dream
the dream we dream together is reality.
User avatar
Posts: 1820
Joined: Tue Jan 18, 2005 12:42 pm
Location: Brisbane, Australia

Return to Toratopes

Who is online

Users browsing this forum: No registered users and 1 guest