Author 
Message 
KL science forum beginner
Joined: 05 Feb 2005
Posts: 12

Posted: Fri Jun 30, 2006 12:13 pm Post subject:
Decomposition of convex polygons



Hi all,
I'm trying to find a way to build an arbitrary convex polygon out of
isoceles triangles (for use in a 3d computer construction), and I've
just about convinced myself this is impossible. Can it be done? What if
the triangles are 'nearly' isoceles (say within 50%)? And of course,
*how* could this be done? The polygon is specified only by the lenght
of all of the sides.
The limitation to isoceles (or nearly such) is built into the computer
system I am working with.
This seems to me to be a relatively interesting geometrical question
which is simply beyond my abilities anymore. I appreciate any help
y'all can give. TIA.
 KL 

Back to top 


Peter Webb science forum Guru Wannabe
Joined: 05 May 2005
Posts: 192

Posted: Fri Jun 30, 2006 3:05 pm Post subject:
Re: Decomposition of convex polygons



"KL" <k.lambda@gmail.com> wrote in message
news:1151669627.649574.303670@75g2000cwc.googlegroups.com...
Quote:  Hi all,
I'm trying to find a way to build an arbitrary convex polygon out of
isoceles triangles (for use in a 3d computer construction), and I've
just about convinced myself this is impossible. Can it be done? What if
the triangles are 'nearly' isoceles (say within 50%)? And of course,
*how* could this be done? The polygon is specified only by the lenght
of all of the sides.
The limitation to isoceles (or nearly such) is built into the computer
system I am working with.
This seems to me to be a relatively interesting geometrical question
which is simply beyond my abilities anymore. I appreciate any help
y'all can give. TIA.
 KL

If I understand you properly (and I probably don't) then it seems very
easily possible.
Pick any point on the surface P and draw a "circle" one unit away. Fill this
with as many isoceles triangles with the pointy end on P and the base
connecting two points on the "circle"  say 6 triangles, so each is
equilateral if it is a flat point. Pick any point on the circle that you
have already used and choose this as your new centre P, and repeat ... you
will tile your surface with icoceles triangles, and "almost" equilateral if
you want. The smaller the size of the unit you use, the finer the modelling,
and the same is true for the number of triangles you fill each circle with,
but I don't know how to optimise the answer. 

Back to top 


KL science forum beginner
Joined: 05 Feb 2005
Posts: 12

Posted: Fri Jun 30, 2006 6:01 pm Post subject:
Re: Decomposition of convex polygons



Peter Webb wrote:
Quote:  "KL" <k.lambda@gmail.com> wrote in message
news:1151669627.649574.303670@75g2000cwc.googlegroups.com...
Hi all,
I'm trying to find a way to build an arbitrary convex polygon out of
isoceles triangles
If I understand you properly (and I probably don't) then it seems very
easily possible.
Pick any point on the surface P and draw a "circle" one unit away. Fill this
with as many isoceles triangles with the pointy end on P and the base
connecting two points on the "circle"  say 6 triangles, so each is
equilateral if it is a flat point. Pick any point on the circle that you
have already used and choose this as your new centre P, and repeat ...

Sorry. Say I want to contruct an irregular pentagon out of isoceles
tris I don't see how your method would do it.
perhaps I should have been more concrete initially...
 KL 

Back to top 


jasen science forum beginner
Joined: 28 Jun 2006
Posts: 16

Posted: Fri Jun 30, 2006 10:07 pm Post subject:
Re: Decomposition of convex polygons



On 20060630, KL <k.lambda@gmail.com> wrote:
Quote:  Hi all,
I'm trying to find a way to build an arbitrary convex polygon out of
isoceles triangles (for use in a 3d computer construction), and I've
just about convinced myself this is impossible. Can it be done? What if
the triangles are 'nearly' isoceles (say within 50%)? And of course,
*how* could this be done? The polygon is specified only by the lenght
of all of the sides.

any polygon can be constructed of arbitrary triangles, so your problem
reduces to constructing an arbitrary triangle from isosceles triangles,
any triangle can be constructed from two rightangle triangles,
(divide it from the longest side to the other corner at a rightangle to the
longest side)
and any right triangle can be constructed from two isosceles
triangles, divide it from the centre of longest side (hypoteneuse) to the
rightangle.
is that proof enough ?

Bye.
Jasen 

Back to top 


Patrick Hamlyn science forum beginner
Joined: 03 May 2005
Posts: 45

Posted: Sat Jul 01, 2006 4:22 am Post subject:
Re: Decomposition of convex polygons



jasen <jasen@free.net.nz> wrote:
Quote:  On 20060630, KL <k.lambda@gmail.com> wrote:
Hi all,
I'm trying to find a way to build an arbitrary convex polygon out of
isoceles triangles (for use in a 3d computer construction), and I've
just about convinced myself this is impossible. Can it be done? What if
the triangles are 'nearly' isoceles (say within 50%)? And of course,
*how* could this be done? The polygon is specified only by the lenght
of all of the sides.
any polygon can be constructed of arbitrary triangles, so your problem
reduces to constructing an arbitrary triangle from isosceles triangles,
any triangle can be constructed from two rightangle triangles,
(divide it from the longest side to the other corner at a rightangle to the
longest side)
and any right triangle can be constructed from two isosceles
triangles, divide it from the centre of longest side (hypoteneuse) to the
rightangle.
is that proof enough ?

That's nice... is there any way of proceeding in similar fashion to prove that
any polygon can be constructed from equilateral triangles? 

Back to top 


jasen science forum beginner
Joined: 28 Jun 2006
Posts: 16

Posted: Sun Jul 02, 2006 2:26 am Post subject:
Re: Decomposition of convex polygons



On 20060701, Patrick Hamlyn <path@multipro.N_OcomSP_AM.au> wrote:
Quote:  jasen <jasen@free.net.nz> wrote:
On 20060630, KL <k.lambda@gmail.com> wrote:
Hi all,
I'm trying to find a way to build an arbitrary convex polygon out of
isoceles triangles (for use in a 3d computer construction), and I've
just about convinced myself this is impossible. Can it be done? What if
the triangles are 'nearly' isoceles (say within 50%)? And of course,
*how* could this be done? The polygon is specified only by the lenght
of all of the sides.
any polygon can be constructed of arbitrary triangles, so your problem
reduces to constructing an arbitrary triangle from isosceles triangles,
any triangle can be constructed from two rightangle triangles,
(divide it from the longest side to the other corner at a rightangle to the
longest side)
and any right triangle can be constructed from two isosceles
triangles, divide it from the centre of longest side (hypoteneuse) to the
rightangle.
is that proof enough ?
That's nice... is there any way of proceeding in similar fashion to prove that
any polygon can be constructed from equilateral triangles?

any polygon, or any convex polygon?
there's a simple recipe to do any convex polgon,
pick any point on the perimeter of the polygon (or even inside it)
and divide the polygon into triangles along lines to every other vertex,
then treat the triangles individually.
I think it can be shown that polygons with concaves can also be done (by
dividing them into convex polygons and then into triangles.
treatment for polygons where lines cross (eg 4 points describing a "bowtie"
or 5 describing a pentagram) depends on what it means when the edges cross.
Bye.
Jasen 

Back to top 


Peter Webb science forum Guru Wannabe
Joined: 05 May 2005
Posts: 192

Posted: Mon Jul 03, 2006 5:47 pm Post subject:
Re: Decomposition of convex polygons



"Patrick Hamlyn" <path@multipro.N_OcomSP_AM.au> wrote in message
news:03uba25cgpdhjs3hhq6qmatain3ttutln9@4ax.com...
Quote:  jasen <jasen@free.net.nz> wrote:
On 20060630, KL <k.lambda@gmail.com> wrote:
Hi all,
I'm trying to find a way to build an arbitrary convex polygon out of
isoceles triangles (for use in a 3d computer construction), and I've
just about convinced myself this is impossible. Can it be done? What if
the triangles are 'nearly' isoceles (say within 50%)? And of course,
*how* could this be done? The polygon is specified only by the lenght
of all of the sides.
any polygon can be constructed of arbitrary triangles, so your problem
reduces to constructing an arbitrary triangle from isosceles triangles,
any triangle can be constructed from two rightangle triangles,
(divide it from the longest side to the other corner at a rightangle to
the
longest side)
and any right triangle can be constructed from two isosceles
triangles, divide it from the centre of longest side (hypoteneuse) to the
rightangle.
is that proof enough ?
That's nice... is there any way of proceeding in similar fashion to prove
that
any polygon can be constructed from equilateral triangles?

No. Every vertex has n equilateral triangles meeting, so must be of angle
n*pi/6. 

Back to top 


Google


Back to top 



The time now is Fri Nov 17, 2017 5:23 pm  All times are GMT

