Author Topic: Spherical knots - Cuboctahedral bend(s)  (Read 17273 times)

xarax

  • Sr. Member
  • *****
  • Posts: 2781
Spherical knots - Cuboctahedral bend(s)
« on: August 08, 2014, 02:01:15 PM »
   A Spherical knot is a knot made by one or more ropes "travelling" on the surface of a sphere.
   One rope makes a stopper, two ropes make a bend ( 2-rope link ).
   The word "sphere" is to be meant with its topological, rather than its geometrical sense. Any 2-D closed surface, topologically equivalent to the sphere, is, in this  sense, "spherical". The elastic surface of a balloon, independently of how much we inflate or we deformed it, locally, by squeezing it, is and remains "spherical", as long as we do not tear it apart.
   All the segments of the ropes cross each other, in pairs, at some points, the "crossing points ". So we should imagine that the crossing points are "points" distributed on the surface of the sphere, and there are two "lines" ( two segments of the one rope, in the case of a stopper, or the two ropes, in the case of a bend ) that cross each other at those "points", going "over" or "under".
   Of course, any knot can be "loosened", and then "flattened", so it will look like the 2-D representations of knots used by the mathematicians. The advantage of a spherical knot is twofold :
   First, the crossing points of a spherical knot can be arranged in such a way that their number is minimal. On the contrary, most of the times, on a loose and flattened knot this will not happen.
  How can we achieve this minimum number of crossing points ? By just copying the diagram of the representation of the corresponding mathematical knot or link shown in a table of knots or links  ( which, by definition, will have the minimum number of crossing points ) and then pasting  / "embedding" it on the surface of a sphere. In other words, it is like we draw such a 2-D representation of our stopper or our bend on a piece of a rubber sheet, and then we wrap a sphere with it, so we cover it completely. A knot that looks like this is spherical, and has the minimum number of crossing points on its surface.
   Second, in a spherical knot all the distances between all the crossing points remain small - because a sphere with a number, say n, of crossing points on its surface, is more compact, spatially, than a planar area with the same number and density of crossing points. By wrapping the sphere with the planar 2_D drawing, the points that use to be on the perimeter of it will come closer together, on the 'rear" surface of the sphere.
   Those two advantages are especially helpful in the case we want to "inflate" the spherical knot, in order to be able to reduce the density of the crossing points on its surface, so the lines that go "over" or "under" those points are more clearly visible. Also, when we have a spherical knot, we can rotate it and look at it from the angle which will show the interweaving of the segments of the rope(s) in a more comprehensible way.

   As an example of a relatively simple, most symmetric spherical knot, the interested reader may try his hand and tie a bend based on a quasi-regular polyhedron ( an Archimedean solid ) with 12 vertices, the cuboctahedron ( see the attached picture ). The blue line represents the continuous path of the one rope, of the first link, on the surface of the one hemisphere, and the red line the continuous path of the other rope, of the second link, on the surface of the other hemisphere. Starting from the "closed" knot, we can "cut" the red and the blue slings at one red and at one blue edge, respectably, and get two pairs of free ends, i.e, get an end-to-ender knot, a bend, like the ones we use. By altering the "over" / "under" relations of the pairs of lines at each crossing point, we can get different bends.
   If we want the most symmetric such knot, the "cuts" should be on symmetrically located edges, and the "over" / "under" relations would form two interlinked mirror symmetric trefoil knots, i.e., two trefoil knots of opposite handedness / chirality - which, when we will "cut" them, they will be transformed in one "cut" left-handed trefoil knot = an overhand knot, and one "cut" right-handed trefoil knot = an underhand knot. The corresponding polyhedron can be seen, animated, in :
    http://upload.wikimedia.org/wikipedia/commons/a/ab/Cuboctahedron.gif
« Last Edit: August 09, 2014, 02:14:25 AM by xarax »
This is not a knot.

struktor

  • Sr. Member
  • *****
  • Posts: 345
Re: Spherical knots - Cuboctahedral bend(s)
« Reply #1 on: August 08, 2014, 08:50:32 PM »
My solutions

« Last Edit: August 13, 2014, 07:26:06 PM by struktor »

xarax

  • Sr. Member
  • *****
  • Posts: 2781
This is not a knot.

xarax

  • Sr. Member
  • *****
  • Posts: 2781
Re: Spherical knots - Cuboctahedral bend(s)
« Reply #3 on: August 09, 2014, 01:16:06 PM »
   Pictures of what I believe is the simplest, most symmetric "Spherical bend", the Cuboctahedral bend.
   As I have said in the first post, every bend can be re-dressed as a "Spherical bend", by re-arranging all the crossing points of its "mathematical", planar representation  ( = the one with the minimum number of crossing points ), now on the shell of sphere, rather than on a plane. However, although by this re-arranging ( which allows them to be readily exposed on the surface of this spherical shell ) all the crossing points become clearly visible, and the distances between them become smaller than in the corresponding "flat", planar representation, other geometrical details may be distorted a lot.
   In a "Spherical bend" which is designed, right from the start, to be spherical, the geometrical distortion of the compact knot is much smaller. The knot is, in a sense, wrapped around an invisible/transparent inner core, and its segments are "travelling" on the surface of the shell of this core, meeting each other at crossing points evenly distributed on it. Conceptually, the knot becomes much simpler, even if its "flattened", loose form is very complex. Moreover, by been able to be represented on a spherical surface, it allows us to revolve , animate or "inflate" it ( and get a geometrically more accurate "exploded" form of it). 
   As just an example of such a symmetric "Spherical bend" I have presented the Cuboctahedral bend. It is not but one overhand and one underhand knots linked together by their free ends : the Standing and the Tail End of the one penetrates the other, so that each one of the two free ends of the one knot goes through one of the two openings of the other knot. The overhand and the underhand knots remain in their Pretzel-like form, so this bend can be considered as yet another Pretzel-to-Pretzel bend, of the many we already have. The two mirror-symmetric Pretzels are minimally interlinked, in the Fisherman s knot way. So, it becomes very easy to memorize and execute the tying of this bend : Tie an overhand near the end of the first rope, then, on the second rope, tie an underhand knot around the Tail End of the first overhand knot, and finally pass the Tail End of the second, underhand knot through the second opening of the first, overhand knot. This tying method has two added advantages :
1. That the overhand and underhand knots can be tightened independently from each other and easily, while they remain at a distance, and be pulled towards each other only afterwards, to slide on their two pair of interpenetrating ends, until they "kiss" each other ( like the two links of a Fisherman s knot ).
2. That the final form of the compact form, even after heavy loading, remains almost identical to the form of the pre-tightened knot - IDEPEDENTLY of which ends are the Standing and which the Tail ends. In the attached picture, I have pre-tightened the knot just a little bit, by pulling the ends of the overhand and of the underhand knot first, and then both pairs of opposite ends, the one after the other.
The interested reader can try the "normal" and the "reverse" forms of this bend - I have absolutely no idea either which is "better", or which is the "normal" and which is the "reversed" bend !  :) 
   This bend looks almost identical to the Forty-Five knot ( M. B23 )(1), the Double S bend ( M. B21. ), the Tweedledum bend ( M. B27 ), the Springy bend (3), Luca s "Highly symmetric overhand bends" (4), and various other Pretzel-to-Pretzel bends. Also, without taking into account the over/under relations, it is similar to the Hexagram bend (5), in which the two Pretzels are maximally interweaved - and, for that reason, it is difficult to tie, and should be considered a decorative bend. On the contrary, the Cuboctahedral bend can be tied very easily and quickly, is conceptually simple and easily inspected, so it can serve as a practical bend.

1. http://igkt.net/sm/index.php?topic=4771.msg30982#msg30982
2. http://igkt.net/sm/index.php?topic=4771.msg31002#msg31002
3. http://igkt.net/sm/index.php?topic=4188.msg25561#msg25561
4. http://igkt.net/sm/index.php?topic=4705.msg30503#msg30503
5. http://igkt.net/sm/index.php?topic=4215.msg25807#msg25807
    http://igkt.net/sm/index.php?topic=3671.msg21234#msg21234
« Last Edit: August 09, 2014, 01:39:43 PM by xarax »
This is not a knot.

struktor

  • Sr. Member
  • *****
  • Posts: 345
Re: Spherical knots - Cuboctahedral bend(s)
« Reply #4 on: August 10, 2014, 01:01:44 PM »
« Last Edit: September 07, 2014, 11:02:29 AM by struktor »

xarax

  • Sr. Member
  • *****
  • Posts: 2781
Re: Spherical knots - Cuboctahedral bend(s)
« Reply #5 on: August 10, 2014, 01:27:05 PM »
   Now you are talking ! You had not shown THAT 2-rope bend in your previous post, had you ?   :)
   Indeed, if we are allowed to replace the simple, X - kind crossing points ( where the two ropes meet each other only at one point )( ABoK#34 ), with more "twisted" "crossing points" ( better, "crossing areas", where the two ropes are twisted around each other, they form a part of a double helix )( ABoK#35, #36 ), like the crossing areas you show in your last image, we get more spherical knots.
   In that case, where the crossings are more complex, we can use a less complex 4-valent polyhedron, the octahedron. I had not explored what one can tie, by using the octahedron and making the two ropes of the two links twist around each other at its 6 crossing points/areas. Perhaps you can tie some of those bends for us, and present them in this thread.
« Last Edit: August 10, 2014, 10:45:16 PM by xarax »
This is not a knot.

xarax

  • Sr. Member
  • *****
  • Posts: 2781
Re: Spherical knots - Cuboctahedral bend(s)
« Reply #6 on: August 10, 2014, 02:07:47 PM »
   It had just occurred to me that, if we allow some edges of the spherical / polyhedral skeletons to represent pahts of two lines ( of a double line ), rather than paths of single lines, we can utilize 3-valent regular or quasi-regular polyhedra : the cube, and, why not, the tetrahedron ! We take a tetrahedron, and retrace its edges with our two lines, so that some ( or all ) of the 6 edges will be retraced by two lines. Allowing those lines to twist around each other / embrace each other at the 4 vertices of the tetrahedron, its "crossing areas", which bends do we get ?
   So, take a transparent ball, mark the vertices of the polyhedron you want to use on its surface, nail your beloved pins on them, and start tying symmetric spherical knots !  :) I know that you will not be restricted to the four simplest regular polyhedra I had mentioned, but those are the ones which may generate something more than a nice decorative knot - a nice simple, practical knot !  :)
« Last Edit: August 10, 2014, 10:46:33 PM by xarax »
This is not a knot.

xarax

  • Sr. Member
  • *****
  • Posts: 2781
Re: Spherical knots - Cuboctahedral bend(s)
« Reply #7 on: August 11, 2014, 01:19:44 PM »
   The Cuboctahedral bend shown in Repl#3, is made from an overhand and an underhand knot, joined together in the Fisherman s knot way : the ends of the knot of the one link penetrate the openings of the knot of the other, so, when the Standing Ends are pulled, the two knots are pushed towards each other.
   What happens if the two links are both overhand or underhand knots ? The result is a topologically slightly, but geometrically much different knot. Now we have a not so symmetric knot, where we can decide with more certainty, based on the different way the two links contact each other, to which ends we should assign the role of the Standing and to which the role of the Tail ends. ( See the bend shown in attached picture ). It is interesting that such a "small" change ( in the handedness of the one link, and nothing else ) is able to generate so different a knot.
This is not a knot.

struktor

  • Sr. Member
  • *****
  • Posts: 345
Re: Spherical knots - Cuboctahedral bend(s)
« Reply #8 on: August 12, 2014, 08:28:18 PM »
Topology  and  THK


« Last Edit: August 13, 2014, 07:24:35 PM by struktor »

xarax

  • Sr. Member
  • *****
  • Posts: 2781
Re: Spherical knots - Cuboctahedral bend(s)
« Reply #9 on: August 12, 2014, 09:05:33 PM »
   All those rosettes can be considered as loose and "flattened" regular cylindrical knots - because they have only one main axis of symmetry/anti-symmetry, the axis of the cylinder. In the regular spherical knots we have more axes of symmetry, each going through the centre of each regular face and the centre of the sphere. Using the regular or quasi-regular polyhedra helps us generate "simpler" knots, I believe.
This is not a knot.

struktor

  • Sr. Member
  • *****
  • Posts: 345
Re: Spherical knots - Cuboctahedral bend(s)
« Reply #10 on: August 15, 2014, 06:06:55 PM »
« Last Edit: September 07, 2014, 09:02:11 AM by struktor »

xarax

  • Sr. Member
  • *****
  • Posts: 2781
Re: Spherical knots - Cuboctahedral bend(s)
« Reply #11 on: August 16, 2014, 03:02:11 AM »
   In the "spherical" representation of the two-strand knot ( the animated one ), the shape and colour symmetries are manifested in a much more clear way - than in the "flattened", rosette-like one. In general, following paths on the surface of the sphere defined by the edges of inscribed regular and quasi-regular polyhedra, is a more efficient way to generate symmetric spherical, and not only cylindrical knots.
   If we remain in the realm of simple X-like crossings, depending on how the ends are crossed as they enter into / exit from the opposed same-colour vertices, the generated octahedral bend can be composed by two interlocked overhand knots, or by two linked Clove hitches (1), or by two linked Backhanded hitches (2) ( See the attached picture ). So, provided we do not twist the strands around each other at the crossing points, the method does not generate any new bend - it seems that the octahedron is too simple for that.

1. http://igkt.net/sm/index.php?topic=1919.msg16425#msg16425
    http://igkt.net/sm/index.php?topic=1919.msg16522#msg16522
    http://igkt.net/sm/index.php?topic=1919.msg16869#msg16869
2. http://igkt.net/sm/index.php?topic=3791.msg22185#msg22185
« Last Edit: August 22, 2014, 08:33:47 AM by xarax »
This is not a knot.

xarax

  • Sr. Member
  • *****
  • Posts: 2781
Re: Spherical knots - Cuboctahedral bend(s)
« Reply #12 on: October 12, 2014, 09:10:40 AM »
   The two links of Octahedral bend as well as of the single and double Fisherman s knot "kiss" each other better when they are mirror symmetric : the one knot is "overhand" and the other is "underhand" - so the crests where the one nub is convex match with the troughs where the other is concave. However, in the case of the Octahedral bend, two mirror symmetric links are joined in one face-symmetric knot, while in the case of the Fisherman s knot, the face-symmetric knot is made by two links of the same helicity / handedness, by two overhand or by two underhand knots. The bend we prefer to use, which is tied and inspected more easily, and where the surfaces of the two interpenetrating single or double overhand/underhand knots "kiss" each other better, is not face-symmetric ( See the attached picture ).
« Last Edit: October 14, 2014, 12:52:53 PM by xarax »
This is not a knot.


xarax

  • Sr. Member
  • *****
  • Posts: 2781
Re: Spherical knots - Cuboctahedral bend(s)
« Reply #14 on: October 13, 2014, 08:06:37 PM »
Why spherical polyhedron?

Maybe ellipsoidal polyhedron?
      biconical polyhedron?
      projective polyhedron?

   The word "sphere" is to be meant with its topological, rather than its geometrical sense. Any 2-D closed surface, topologically equivalent to the sphere, is, in this  sense, "spherical". The elastic surface of a balloon, independently of how much we inflate or we deforme it, locally, by squeezing it, is and remains "spherical", as long as we do not tear it apart.
« Last Edit: October 13, 2014, 08:07:13 PM by xarax »
This is not a knot.