![Definition of a Map. Abstracting the definition of a map given in [1]:
A map is a graph drawn on a closed surface in such a way that:
a) the vertices are represented as distinct points on the surface,
b) the edges are represented as curves on the surface intersecting only at the vertices,
c) if we cut the surface along the graph, what remains is a disjoint union of connected components,
called faces, each topologically equivalent to an open disk.
Examples: a monogon (i.e., a single vertex with a self-loop) is a map when drawn on the sphere; on the
torus it can be drawn several different ways, but none succeeds in being a map because one of the faces
will fail to be a topological disk. The skeleton graph of a tetrahedron is a map when drawn on the sphere
without crossings—and also a map if drawn on the torus in a particular way (see Figure 1.) In computer
science terms: if a graph encodes the topology of the surface it is drawn on, then it's a map.
A practical use of maps is to mathematically describe a surface we are interested in making or decorating.
Map Operations. Map operations are procedures that convert one map into another [2-3]. Map
operations are known under various guises in different fields, most notably in Conway's polyhedron
notation [4], in computer graphics, where they appear as various subdivision schemes [5-8], and in
geodesic dome design, where they appear as various geodesic breakdown schemes [9].
Many map operations can operate indiscriminately on map faces that are monogons, digons, or n-gons.
Many map operations (namely, the achiral ones) work just as well on surfaces that are orientable and
nonorientable, and surfaces of any topological genus. This wide scope of application is why we should
indeed prefer to speak of maps, not limiting ourselves to polyhedra or computer graphics surface models.
Some map operations are fundamental: identity, dual, and reflect are of first importance (reflect [4],
unfortunately, is one that cannot be accomplished with truchet tiles.)
Some map operations are composite, meaning that they can be accomplished by performing a sequence of
other map operations in a specified order (note that map operations are not generally commutative.) But,
since truchet tiles cannot compose map operations—nor can the visual imagination do this very easily—
some composite map operations are presented here.
Names and Notation. I have strayed from Conway’s operation names only in the case of medial and
radial, terms well established in the graph theory literature. Conway’s concise notation system can be a
little cryptic for the novice, and, in any case, with twenty-one (and counting) map operations, a suitably
mnemonic mapping to the 26 letters of the alphabet is beyond reach. Instead, I adopt here the two-letter
function notation common in the map operations literature. For the names of map operations not in
Conway, I have adopted Hart’s propellor and reflect [4], and elsewhere followed precedents in the map
operations literature, but I have rejected subscripts and numerals, and insisted that “action” names be in
verb form. Two operations well-known in the computer graphics literature, have been dubbed with the
first two letters of the names of the authors most associated with each [5-6]. A composite operation which
is particularly useful in ring weaving has been dubbed ring.
Having abbreviations of operation names allows one to write out composite operations and identities in
function notation. For example, the fact that a map and its dual yield the same medial is expressed:
Me(M) = Me(Du(M)).](https://image.slidesharecdn.com/extrawayspaper-110622100245-phpapp02/85/Extra-ways-to-see-An-Artist-s-Guide-to-Map-Operations-2-320.jpg)
![Practical Applications
Plain Weaving. A common problem in weaving is imagining the reverse face of the work. For example, a
crossing of a vertical thread with a horizontal thread will—as we might expect—look different on the
reverse face, but a crossing of two oblique threads looks the same! Confusion multiplies when one tries to
weave a non-orientable surface such as a mobius strip. The surest resort is a technique known in knot
theory as Tait’s coloring of link diagrams [10], but first suggested in relation to weaving by Snelson [11]
and fully developed by Roelofs [12].
In plain-weaving [13], which is weaving where every element passes over and under in strict alternation,
and no three elements cross at the same point, each opening in the fabric is surrounded by a helical
arrangement of weaving elements. Any such helix can be assigned a left- or right-handedness using the
same rules as for an ordinary screw. This handedness is intrinsic, it does not change with the way we look
at the opening. What is more, in plain-weaving each fabric opening shares edges exclusively with
openings of the opposite handedness. Such a spatial arrangement is mathematically termed a proper face
2-coloring, or chess coloring [14]. Also, since elements of a plain-weaving cross strictly two-at-a-time,
the map of a plain-weaving is 4-regular. Thus, more useful than a pictorial representation of a plain
weaving, is a chess-colored, 4-regular map for it. If such a map is imagined to be inked “right through the
paper,” it becomes a reliable guide to both faces of the weaving.
Chess black Chess white
Figure 5: Monogon representations of chess-colored medial tiles for plain-weaving.
The map operation medial always yields a four-regular, chess-colorable map. In fact, a map M is a medial
map of some map N if and only if M is four-regular and its faces can be chess-colored [15]. Medial is the
refore the whole story for plain-weaving. All plain-woven baskets correspond to medial maps, and any
given map can be converted into a design for plain-woven basket via the medial operation and a chess
coloring. Medial can be understood as the projection onto the surface of the topological twisting operation
in [13].
Ring Weaving. Ring weaving, a.k.a. chain mail, is often a variety of plain-weaving. A design for the
commonest plain-weave chain mail is generated from any map by chess coloring the map resulting from
the map operation ring,
Ri(M) = Me(Su(M)).](https://image.slidesharecdn.com/extrawayspaper-110622100245-phpapp02/85/Extra-ways-to-see-An-Artist-s-Guide-to-Map-Operations-8-320.jpg)
![Coloring Operations. A suitable chess-coloring can be pre-marked on the truchet tile for a map operation
since such a coloring is essentially unique. If a map is chess-colorable, then its dual map is bipartite,
meaning that its vertices can be colored black and white in such a way that no two vertices of the same
color are linked by an edge. Bipartite maps are important in the study of Riemann surfaces, where the
map operation subdivide is used to generate bipartite maps which are colored according to the convention
that the “old” vertices are colored black. It seems useful to generalize this coloring convention. Define a
(meta) map operation, bipartite black, Bb(Xx(M)), as the operation that generates the bipartite-colored
map (if such exists) of the map Xx(M), using the convention that all the “old” vertices in Xx(M) are to be
black. Bipartite white, Bw(Xy(M)), generates the inverse coloring. Dually, define chess black,
Cb(Yy(M)), as the operation that generates the chess-colored map (if such exists) of Yy(M), using the
convention that all the “old” faces of Yy(M), are colored black. Chess white, Cw(Yy(M)), generates the
inverse coloring. Figure 5 shows the action of chess black and chess white on medial.
Tensegrity. Pugh [16] describes three categories of tensegrity structures, diamond, zig-zag, and circuit.
All three can be associated with map operations. The diamond tensegrity pattern is isomorphic to snub.
The compression elements, or struts, of the tensegrity structure, correspond to the chiral edge of snub; the
remaining non-chiral edges (which, by themselves, could equally well have been generated by the map
operation expand,) correspond to the tension elements, or tendons, of the tensegrity structure. A monogon
tiling for snub with the strut indicated by a thicker line is shown in Figure 7a. When the base map is a
triangle on the sphere, this map operation produces the famous 3-strut, 9-tendon, T-prism. When the base
map is a tetrahedron, the result is the equally well-known 6-strut, 24-tendon, expanded octahedron.
The situation is a bit messier for the zig-zag and circuit tensegrity patterns. These patterns have elements
that crossover each other without actually touching—thus breaking one of the rules for maps. A stratagem
is to portray these patterns with the (narrow) tendons overlying the (wide) struts—not indicating that they
pass in front, but rather that they may pass either in front or behind—as the curvature of the surface
dictates. Such “floating” tendons and struts are non-physical, but the problem of deciding whether
tendons or struts pass in front is fully determined, in practice, by the spatial coordinates of their
endpoints. (If the surface the tensegrity structure conforms to is sufficiently curved, strut-tendon and strut-
strut collisions are avoided.) In circuit, the tendons and struts both align with the edges of medial, but
struts join in the common vertex only on alternate passes (see Figure 7b.) In zig-zag, tendons align with
the edges of truncate, while struts align with an additional chiral edge which crosses over a tendon at the
equator (see Figure 7c.)
Diamond Zig-zag Circuit
Figure 7: Monogon representations of tensegrity patterns.
Texture Mapping Examples of practical uses of map operations are shown in Figure 8. These images
were realized by texture-mapping a triplet of kis-truchet tiles (in effect, a truchet tile composing a full
equilateral triangle) onto 3-D models having all-triangle meshes using Processing [17].](https://image.slidesharecdn.com/extrawayspaper-110622100245-phpapp02/85/Extra-ways-to-see-An-Artist-s-Guide-to-Map-Operations-9-320.jpg)
![References
[1] S. K. Lando, A. K. Zvonkin, and D. B. Zagier, Graphs on Surfaces and Their Applications,
Encyclopedia of Mathematical Sciences, Low Dimensional Topology, p. 28, Springer-Verlag, 2004.
[2] T. Pisanski and M. Randic, Bridges between geometry and graph theory, Geometry at Work, C. A.
Gorini, ed., pp. 174-194, Mathematical Association of America, 2000.
[3] M. V. Diudea, M. Stefu, P. E. John, A. Graovac, Generalized operations on maps, Croatica Chemica
Acta, CCACAA 79 (3) 355-362, 2006.
[4] G. W. Hart, Conway notation for polyhedra, www.georgehart.com/virtual-polyhedra/
conway_notation.html
[5] C. Loop, Smooth Subdivision Surfaces Based on Triangles. Master’s thesis, University of Utah, 1987.
[6] L. Kobbelt, √3-Subdivision, Proc. ACM Siggraph 2000.
[7] J. Peters and U. Reif, The simplest subdivision scheme for smoothing polyhedra, Purdue University,
1996
[8] P. Oswald and P. Schroder, Composite primal/dual √3-subdivision schemes, Computer Aided
Geometric Design, vol. 20 n. 3, pp. 135-164, 2003.
[9] H. Kenner, Geodesic Math and How to Use It, chap. 5, Univ. of California Press, 1976.
[10] D. J. A. Welsh, Complexity: Knots, Colourings and Counting, p. 23, Cambridge Univ. Press, 1993.
[11] K. Snelson, Space frame structure made by 3-D weaving of rod members, U.S. Pat. 6739937, 2004.
[12] R. Roelofs, About weaving and helical holes, Bridges 2010.
[13] E. Akleman, J. Chen, Q. Xing, and J. Gross "Cyclic plain-weaving on polygonal mesh surfaces with
extended graph rotation systems," Proc. ACM Siggraph ’2009.
[14] Deza, M., and Dutour, M., Zigzags and central circuits for 3- or 4-valent plane graphs,
www.liga.ens.fr/~deza/Sem-ZigCc/ZigCcConf.pdf
[15] D. Archdeacon, Representativity, The Handbook of Graph Theory, J. L. Gross and J. Yellen eds.,
chap 7.7, F4, p. 724, CRC Press, 2003.
[16] A. Pugh, An Introduction to Tensegrity, Univ. of California Press, 1976.
[17] www.processing.org](https://image.slidesharecdn.com/extrawayspaper-110622100245-phpapp02/85/Extra-ways-to-see-An-Artist-s-Guide-to-Map-Operations-11-320.jpg)
Extra ways to see: An Artist's Guide to Map Operations
- 1. Extra Ways to See: An Artist’s Guide to Map Operations James Mallos Sculptor Silver Spring, MD, USA Abstract A map operation converts one mathematical model of a surface into another, and can thereby suggest to the artist or artisan different ways of building or decorating a given surface. A truchet tile technique is presented that allows many common map operations to be represented by truchet tiles of triangular shape. These map operations can also be defined pictorially by their action on a monogon embedded in the sphere. Truchet tiles for twenty-one map operations are illustrated along with examples of their use in the design of weaving and tensegrity structures. Introduction Familiarity with map operations can spark the visual imagination. For example, it might be surprising that a map with all triangular faces (think of the edges of an icosahedron projected onto the sphere) can be trivially converted into a map with all quadrilateral faces, or into one with all pentagonal faces, or into a design for a weaving, or into a design for a tensegrity structure. Seemingly alchemical changes like these are effected by map operations. The operations themselves are quite simple, so simple that many can be represented visually by truchet tiles; that is, by specially patterned tiles designed to fit together, side-by- side, to help visualize a larger pattern. This article illustrates truchet tiles for many map operations and gives some examples of their practical use. b f d c e a g Figure 1: Which of these are maps? Only b, c, and f.
- 2. Definition of a Map. Abstracting the definition of a map given in [1]: A map is a graph drawn on a closed surface in such a way that: a) the vertices are represented as distinct points on the surface, b) the edges are represented as curves on the surface intersecting only at the vertices, c) if we cut the surface along the graph, what remains is a disjoint union of connected components, called faces, each topologically equivalent to an open disk. Examples: a monogon (i.e., a single vertex with a self-loop) is a map when drawn on the sphere; on the torus it can be drawn several different ways, but none succeeds in being a map because one of the faces will fail to be a topological disk. The skeleton graph of a tetrahedron is a map when drawn on the sphere without crossings—and also a map if drawn on the torus in a particular way (see Figure 1.) In computer science terms: if a graph encodes the topology of the surface it is drawn on, then it's a map. A practical use of maps is to mathematically describe a surface we are interested in making or decorating. Map Operations. Map operations are procedures that convert one map into another [2-3]. Map operations are known under various guises in different fields, most notably in Conway's polyhedron notation [4], in computer graphics, where they appear as various subdivision schemes [5-8], and in geodesic dome design, where they appear as various geodesic breakdown schemes [9]. Many map operations can operate indiscriminately on map faces that are monogons, digons, or n-gons. Many map operations (namely, the achiral ones) work just as well on surfaces that are orientable and nonorientable, and surfaces of any topological genus. This wide scope of application is why we should indeed prefer to speak of maps, not limiting ourselves to polyhedra or computer graphics surface models. Some map operations are fundamental: identity, dual, and reflect are of first importance (reflect [4], unfortunately, is one that cannot be accomplished with truchet tiles.) Some map operations are composite, meaning that they can be accomplished by performing a sequence of other map operations in a specified order (note that map operations are not generally commutative.) But, since truchet tiles cannot compose map operations—nor can the visual imagination do this very easily— some composite map operations are presented here. Names and Notation. I have strayed from Conway’s operation names only in the case of medial and radial, terms well established in the graph theory literature. Conway’s concise notation system can be a little cryptic for the novice, and, in any case, with twenty-one (and counting) map operations, a suitably mnemonic mapping to the 26 letters of the alphabet is beyond reach. Instead, I adopt here the two-letter function notation common in the map operations literature. For the names of map operations not in Conway, I have adopted Hart’s propellor and reflect [4], and elsewhere followed precedents in the map operations literature, but I have rejected subscripts and numerals, and insisted that “action” names be in verb form. Two operations well-known in the computer graphics literature, have been dubbed with the first two letters of the names of the authors most associated with each [5-6]. A composite operation which is particularly useful in ring weaving has been dubbed ring. Having abbreviations of operation names allows one to write out composite operations and identities in function notation. For example, the fact that a map and its dual yield the same medial is expressed: Me(M) = Me(Du(M)).
- 3. Order, Type, and Chirality. Map operations can be ranked in orders according to the number of edges they place in the resultant map when the base map is a monogon drawn on the sphere. For example, identity, dual, and reflect make up the first order, each placing a single edge in the resulting spherical map. Subdivide, parallel, radial, and medial make up the second order, each placing two edges in the resulting spherical map... and so on. Map operations can also be typed according to whether, like identity, they include the original vertex of the monogon base-map, or, like dual, they delete it. The former will be called i-type, and the latter d-type. Map operations that are chiral come in left-handed and right-handed varieties (thus they cannot be used on a surface that cannot be oriented, such as a mobius strip.) All four of the chiral map operations in the present collection are fifth order. Figure 2: Stretchy triangular tiles can stretch to tile digons (above) and monogons (below.) Representing Map Operations with Truchet Tiles Stretchy Tiles. The truchet tiles contemplated in this paper are stretchy. That is, we will imagine that a patterned triangular tile can be stretched to conform to any size or shape of triangle. A vertex angle can be stretched open to 180 degrees—enabling two such tiles to tile a digon (see Figure 2.) Even further, a vertex angle can be stretched open to a full 360 degrees, enabling it to tile a monogon by itself. Note that the stretched triangles still have three sides in these tilings—they are topologically still triangles. Though paper or ceramic tiles cannot stretch in this way, it is possible to mathematically define a mapping (texture-mapping) carrying a pattern or texture from an equilateral triangle onto a half-circle or full circle. If the truchet pattern represents fiber elements (weaving) or building elements (tensegrity) the pattern joins across tile boundaries are equally valid when the the tile has been stretched in this way.
- 4. Figure 3: Kis on an arbitrary map. Every face in the resultant map is a triangle. Kis First. To effect map operations with truchet tiles that are all of the same shape, we need to give priority to a particular map operation. Conway's name for it is kis. A truchet tile that pictorially defines kis will be introduced later, but in every case we will need to perform kis algorithmically as a preliminary to laying down tiles. The preliminary use of kis erases the base map (i.e., the map we are performing the operation on,) and replaces it with a set of "chalk" guidelines whose ultimate fate is to lie hidden behind truchet tiles (see Figure 3.) Algorithm for the Preliminary Kis: with pink chalk, mark a new vertex in the center of each face in the base map, with blue chalk, draw edges connecting the new vertex to each of the old vertices surrounding that face, erase-and-redraw, using blue chalk, all of the old vertices and edges of the base map. The kis operation subdivides every face of the base map into triangles, including monogon faces, digon faces, and faces that were triangles to start off with. It thereby eliminates the need for truchet tiles in any other shape: all we need are tiles in the shape of kis triangles. We will refer to these as kis-truchet tiles. Note that kis-truchet tiles are not rotationally symmetric. That is why the new vertices are colored pink: to mark how the kis-truchet tiles should be oriented. Illustrating Kis-Truchet Tiles. There are two ways to illustrate a kis-truchet tile. Drawn as a one-third portion of an equilateral triangle (see Figure 4,) a kis-truchet tile looks familiarly like other truchet tiles. Alternatively, since they are stretchy, a kis-truchet tile can be illustrated tiling a monogon on the sphere. A way to visualize the stretching involved is to cut a kis-truchet triangle out of paper and roll it into a cone with the pink vertex at the apex. Viewing the cone end-on reveals the tile wrapped inside a teardrop- shaped monogon (see Figure 5.)
- 5. Figure 4: A kis-truchet tile for medial shown as 1/3 of a triangle, left, and tiling a monogon, right. The Chart of Map Operations. In the following Chart, kis-truchet tiles for twenty-one map operations are illustrated, both as one-third portions of equilateral triangles and as tilings of a monogon on the sphere. In interpreting the illustration of the monogon tiling, one should imagine that the sphere is wearing the self-loop and vertex of the monogon base-map like a belt and buckle around its equator. Imagine yourself to be standing "buckle-to-buckle" with the sphere looking down on its northern hemisphere. (A glance in the Chart at the monogon representation for identity shows how this looks.) In the chart, all the kis- truchet triangles are shown with the pink vertex pointing down in order to make it easier to mentally roll them into the cones that would correspond to the monogon representation. The sphere also has a southern hemisphere which is tiled by any map operation. If we were to turn the sphere over (still keeping it buckle-to-buckle to us) the tiling of the southern hemisphere would reveal itself to be exactly like that of the northern hemisphere. Alternatively, if we prefer to keep things stationary and view the southern hemisphere through the northern hemisphere with “x-ray vision,” the southern hemisphere looks like a mirror image of the northern hemisphere. In consequence, each point on the equator is directly (left or right) across from a point on the opposite limb of the sphere that is its other half. This perimetral bilateral symmetry, along with the practical needs to produce a connected map and, in general, avoid 1-valent and Figure 5: Kis-truchets rolled into cones with the 2-valent vertices, are the main constraints on useful pink vertex at the apex. map operations.
- 6. Chart of Map Operations Map Operation Monogon Kis-Truchet Triangle Identities Aliases Representation Order-Type Properties Identity Id(M) = M Id() 1i Dual Du(Du(M)) = M Poincare Dual Du() 1d Subdivide Su(M) = Su1, 1-Dimensional Subdivision Du(Pa(Du(M))) Su() Inserts 2-valent 2i vertices in edges. Parallel Pa(M) = Parallelization Du(Su(Du(M))) Pa() 2i Doubles all edges. Radial Ra(M) = Du(Me(M)) Join Ra(M) = Ra(Du(M)) Ra() Result is bipartite & 2i quadrangle-faced. Medial Me(M) = Du(Ra(M)) Ambo, Midpoint Subdivision, Mid-Edge Me(M) = Me(Du(M)) Me() Result is 4-regular & Subdivision 2d chess-colorable. Kis Ki(M) = P3, Su2, 2-Dimensional Subdivision, Stellation, Du(Tr(Du(M))) Ki() Result is triangle- Omnicapping 3i faced. Truncate Tr(M) = Truncation Du(Ki(Du(M))) Tr() 3d Result is 3-regular. Leapfrog Le(M) = Du(Ki(M)) Tripling, Dual √3 Trisection Le(M) = Tr(Du(M)) Le() 3d Result is 3-regular Chamfer Q, Quadrupling Ch() 4i
- 7. Ortho Or(M) = Ra(Ra(M)) P4, Tetrangulation, Edge Bisection, Catmull-Clark Or(M) = Du(Ex(M)) Or() Result is bipartite & Subdivision, Primal Quadrilateral Quadrisection 4i quadrilateral-faced. Lo Lo(M) = Ko(Du(M)) 2ν Geodesic Breakdown, Loop Subdivision, Primal Lo() Superposes Me(M) Triangle Quadrisection 4i and Id(M). Ko Ko(M) = Lo(Du(M)) Primal √3 Trisection Ko() Superposes Ra(M) 4i and Du(M). Expand Ex(M) = Me(Me(M)) Cantellation, Mrs. Stottʼs Expansion Operation, Ex(M) = Du(Or(M)) Ex() Result is 4-regular & Doo-Sabin Subdivision, Dual Quadrilateral Quadrisection 4d chess-colorable. Ring Ri(M) = Me(Su(M)) Ri() 4d Propellor Pr() 5i chiral Capra Septupling Ca() 5i chiral Ch(M) + chiral edge. Gyro Gy(M) = Du(Sn(M)) P5, Pentangulation Gy() Result is pentagon- 5i chiral faced. Snub Sn(M) = Du(Gy(M)) Sn(Du(M) = Sn(M) Sn() Result is 5-regular. 5d chiral Ex(M) + chiral edge Meta Mt(M) = Ki(Ra(M)) Full Bisection, Barycentric Subdivision, 2-D Mt(M) = Du(Be(M)) Mt() Result is triangle- Subdivision, Dual Triangle Quadrisection 6i faced. Bevel Be(M) = Du(Mt(M)) Omnitruncation Be(M) = Tr(Me(M)) Be() 6d Result is 3-regular.
- 8. Practical Applications Plain Weaving. A common problem in weaving is imagining the reverse face of the work. For example, a crossing of a vertical thread with a horizontal thread will—as we might expect—look different on the reverse face, but a crossing of two oblique threads looks the same! Confusion multiplies when one tries to weave a non-orientable surface such as a mobius strip. The surest resort is a technique known in knot theory as Tait’s coloring of link diagrams [10], but first suggested in relation to weaving by Snelson [11] and fully developed by Roelofs [12]. In plain-weaving [13], which is weaving where every element passes over and under in strict alternation, and no three elements cross at the same point, each opening in the fabric is surrounded by a helical arrangement of weaving elements. Any such helix can be assigned a left- or right-handedness using the same rules as for an ordinary screw. This handedness is intrinsic, it does not change with the way we look at the opening. What is more, in plain-weaving each fabric opening shares edges exclusively with openings of the opposite handedness. Such a spatial arrangement is mathematically termed a proper face 2-coloring, or chess coloring [14]. Also, since elements of a plain-weaving cross strictly two-at-a-time, the map of a plain-weaving is 4-regular. Thus, more useful than a pictorial representation of a plain weaving, is a chess-colored, 4-regular map for it. If such a map is imagined to be inked “right through the paper,” it becomes a reliable guide to both faces of the weaving. Chess black Chess white Figure 5: Monogon representations of chess-colored medial tiles for plain-weaving. The map operation medial always yields a four-regular, chess-colorable map. In fact, a map M is a medial map of some map N if and only if M is four-regular and its faces can be chess-colored [15]. Medial is the refore the whole story for plain-weaving. All plain-woven baskets correspond to medial maps, and any given map can be converted into a design for plain-woven basket via the medial operation and a chess coloring. Medial can be understood as the projection onto the surface of the topological twisting operation in [13]. Ring Weaving. Ring weaving, a.k.a. chain mail, is often a variety of plain-weaving. A design for the commonest plain-weave chain mail is generated from any map by chess coloring the map resulting from the map operation ring, Ri(M) = Me(Su(M)).
- 9. Coloring Operations. A suitable chess-coloring can be pre-marked on the truchet tile for a map operation since such a coloring is essentially unique. If a map is chess-colorable, then its dual map is bipartite, meaning that its vertices can be colored black and white in such a way that no two vertices of the same color are linked by an edge. Bipartite maps are important in the study of Riemann surfaces, where the map operation subdivide is used to generate bipartite maps which are colored according to the convention that the “old” vertices are colored black. It seems useful to generalize this coloring convention. Define a (meta) map operation, bipartite black, Bb(Xx(M)), as the operation that generates the bipartite-colored map (if such exists) of the map Xx(M), using the convention that all the “old” vertices in Xx(M) are to be black. Bipartite white, Bw(Xy(M)), generates the inverse coloring. Dually, define chess black, Cb(Yy(M)), as the operation that generates the chess-colored map (if such exists) of Yy(M), using the convention that all the “old” faces of Yy(M), are colored black. Chess white, Cw(Yy(M)), generates the inverse coloring. Figure 5 shows the action of chess black and chess white on medial. Tensegrity. Pugh [16] describes three categories of tensegrity structures, diamond, zig-zag, and circuit. All three can be associated with map operations. The diamond tensegrity pattern is isomorphic to snub. The compression elements, or struts, of the tensegrity structure, correspond to the chiral edge of snub; the remaining non-chiral edges (which, by themselves, could equally well have been generated by the map operation expand,) correspond to the tension elements, or tendons, of the tensegrity structure. A monogon tiling for snub with the strut indicated by a thicker line is shown in Figure 7a. When the base map is a triangle on the sphere, this map operation produces the famous 3-strut, 9-tendon, T-prism. When the base map is a tetrahedron, the result is the equally well-known 6-strut, 24-tendon, expanded octahedron. The situation is a bit messier for the zig-zag and circuit tensegrity patterns. These patterns have elements that crossover each other without actually touching—thus breaking one of the rules for maps. A stratagem is to portray these patterns with the (narrow) tendons overlying the (wide) struts—not indicating that they pass in front, but rather that they may pass either in front or behind—as the curvature of the surface dictates. Such “floating” tendons and struts are non-physical, but the problem of deciding whether tendons or struts pass in front is fully determined, in practice, by the spatial coordinates of their endpoints. (If the surface the tensegrity structure conforms to is sufficiently curved, strut-tendon and strut- strut collisions are avoided.) In circuit, the tendons and struts both align with the edges of medial, but struts join in the common vertex only on alternate passes (see Figure 7b.) In zig-zag, tendons align with the edges of truncate, while struts align with an additional chiral edge which crosses over a tendon at the equator (see Figure 7c.) Diamond Zig-zag Circuit Figure 7: Monogon representations of tensegrity patterns. Texture Mapping Examples of practical uses of map operations are shown in Figure 8. These images were realized by texture-mapping a triplet of kis-truchet tiles (in effect, a truchet tile composing a full equilateral triangle) onto 3-D models having all-triangle meshes using Processing [17].
- 10. M Me(M) ➯ Cb(Me(M)) Plain-weaving ➯ Sn(M) Diamond tensegrity Figure 8: Practical applications of map operations. Models courtesy of INRIA via the Aim@Shape Shape Repository.
- 11. References [1] S. K. Lando, A. K. Zvonkin, and D. B. Zagier, Graphs on Surfaces and Their Applications, Encyclopedia of Mathematical Sciences, Low Dimensional Topology, p. 28, Springer-Verlag, 2004. [2] T. Pisanski and M. Randic, Bridges between geometry and graph theory, Geometry at Work, C. A. Gorini, ed., pp. 174-194, Mathematical Association of America, 2000. [3] M. V. Diudea, M. Stefu, P. E. John, A. Graovac, Generalized operations on maps, Croatica Chemica Acta, CCACAA 79 (3) 355-362, 2006. [4] G. W. Hart, Conway notation for polyhedra, www.georgehart.com/virtual-polyhedra/ conway_notation.html [5] C. Loop, Smooth Subdivision Surfaces Based on Triangles. Master’s thesis, University of Utah, 1987. [6] L. Kobbelt, √3-Subdivision, Proc. ACM Siggraph 2000. [7] J. Peters and U. Reif, The simplest subdivision scheme for smoothing polyhedra, Purdue University, 1996 [8] P. Oswald and P. Schroder, Composite primal/dual √3-subdivision schemes, Computer Aided Geometric Design, vol. 20 n. 3, pp. 135-164, 2003. [9] H. Kenner, Geodesic Math and How to Use It, chap. 5, Univ. of California Press, 1976. [10] D. J. A. Welsh, Complexity: Knots, Colourings and Counting, p. 23, Cambridge Univ. Press, 1993. [11] K. Snelson, Space frame structure made by 3-D weaving of rod members, U.S. Pat. 6739937, 2004. [12] R. Roelofs, About weaving and helical holes, Bridges 2010. [13] E. Akleman, J. Chen, Q. Xing, and J. Gross "Cyclic plain-weaving on polygonal mesh surfaces with extended graph rotation systems," Proc. ACM Siggraph ’2009. [14] Deza, M., and Dutour, M., Zigzags and central circuits for 3- or 4-valent plane graphs, www.liga.ens.fr/~deza/Sem-ZigCc/ZigCcConf.pdf [15] D. Archdeacon, Representativity, The Handbook of Graph Theory, J. L. Gross and J. Yellen eds., chap 7.7, F4, p. 724, CRC Press, 2003. [16] A. Pugh, An Introduction to Tensegrity, Univ. of California Press, 1976. [17] www.processing.org